Welcome

About this hub

Open any section at any time — Neurocognitive disorders, Clinical anatomy and neuroradiology, Capacity assessment, Fitness to drive, Clinic companion, or Education. Rest on any of them to see what is inside it and jump straight to a section. The logo runs the sidebar itself: click it to fold the list down to icons, and click it again to drop the full list back out over the page. On a phone the same logo drops the list down from the title bar. Search covers every page and the text inside the embedded tools; press / to reach it from anywhere.

Rhymie Yusoff

Rhymie is a Geriatric Medicine Specialist (Consultant Geriatrician) with a longstanding interest in artificial intelligence, technology, and medical education. This hub grew out of that interest — an attempt to bring the reference material, bedside tools, and teaching resources used day to day in geriatric and rehabilitation practice into one searchable place, built and kept up to date with the help of AI-assisted tools.

Geriatric medicine Artificial intelligence Health technology Medical education

Disclaimer

This hub is for guidance and educational purposes only. It is not a clinical guideline, protocol, or institutional policy, and it must not be used as a substitute for one. Always follow your local institution's approved guidelines and protocols, consult the current primary sources, and apply individual clinical judgement for each patient.
Neurocognitive disorder pathway

Neurocognitive disorders

A DSM-5-TR–aligned pathway for formulating major and mild neurocognitive disorder in geriatric practice, plus the bedside tools built alongside it.

Formulation to diagnosis of neurocognitive disorder

The four DSM-5-TR criteria, worked through in order — each step below expands into the supporting detail. Once all four are satisfied, move to the Detailed guide tab for testing, subtype interpretation, staging, and management.

A

Symptoms across a cognitive domain

Decline in one or more of six defined cognitive domains — not just memory — evidenced by concern from the patient, an informant, or the clinician, and objective testing where possible. SAMPLES keeps all six in mind.

SAMPLES — the six DSM-5-TR domains
LetterStands forMaps to
SSignificant impairment in independenceCriterion B (Major NCD only) — not a cognitive domain itself
AComplex attentionSustained/divided/selective attention, processing speed
MMemory and learningImmediate span, recent memory, semantic/autobiographical memory
PPerceptual-motor functionVisual perception, visuoconstruction, praxis, gnosis
LLanguageNaming, word-finding, fluency, grammar, comprehension
EExecutive dysfunctionPlanning, working memory, inhibition, mental flexibility
SSocial cognitionEmotion recognition, theory of mind, behavioural regulation
B

Loss of independence — Major vs. Mild NCD

Whether the decline interferes with independence in everyday activities is what separates Major NCD (dementia) from Mild NCD (MCI) — not the severity of the test score.

Mild NCD (MCI)Major NCD (dementia)
Decline is present but independence in instrumental activities of daily living (IADLs) is preserved — even if it now takes more time, more effort, or compensatory strategies such as lists and reminders.Decline interferes with independence — the person needs assistance with instrumental (finances, medications, transport) or basic activities of daily living.
Why this distinction matters
MCI (Mild NCD) vs. dementia (Major NCD): it is functional independence, not the test score, that is decisive under Criterion B. A low score with preserved independence (more effort/compensation only) = Mild NCD; a similar score with genuine need for IADL assistance = Major NCD. DSM-5-TR's Mild NCD and the research literature's "MCI" are the same construct — either label is acceptable for documentation as long as the aetiological subtype is specified.
C

Exclude delirium

Severity and aetiology should not be finalised until delirium has been excluded or has resolved — screen with 4AT (or CAM/CAM-ICU in hospital) at first contact.

4AT and CAM scoring
ToolItems scoredPositive result
4ATAlertness; AMT4 (age, DOB, place, year); Attention (months backwards); acute change/fluctuation over 2 weeks≥4 = probable delirium; 1–3 = possible cognitive impairment; 0 = delirium unlikely (but doesn't fully exclude it if attention couldn't be tested)
CAMFeature 1: acute onset/fluctuating course; Feature 2: inattention; Feature 3: disorganised thinking; Feature 4: altered consciousnessPositive = Feature 1 AND 2, plus either 3 or 4
D

Exclude mimics

The decline should not be better explained by another mental disorder — screen for the common mimics before finalising a neurocognitive diagnosis.

Common mimics to screen for
  • Depression / pseudodementia
    Often subacute onset, prominent "I don't know" answers, effort-dependent testing, mood symptoms preceding cognitive complaints — improves with treatment of the mood disorder.
  • Other psychiatric disorders
    Anxiety, psychotic disorders, and severe distress can all impair attention and processing speed on bedside testing.
  • Substance or medication effects
    Anticholinergics, benzodiazepines, opioids, alcohol — check the temporal relationship and consider a deprescribing trial.
  • Uncorrected sensory impairment
    Hearing or visual loss can masquerade as cognitive impairment on bedside testing.
  • Delirium not yet resolved
    Revisit Criterion C if the picture is fluctuating or acute.
  • Reversible metabolic / structural causes
    Thyroid dysfunction, B12/folate deficiency, normal pressure hydrocephalus, subdural haematoma, and other causes on the standard dementia screen.
All four criteria met → formulate the neurocognitive disorder and its likely aetiological subtype. The Detailed guide tab covers screening tools, subtype pattern recognition, staging, and management.

The consultation in the order it actually runs — ten steps from the first history question to the advance care plan. Each step opens on click. The four DSM-5-TR criteria in summary form live on the Flowchart tab; steps 5, 6 and 9 below are where they get worked through properly. The full 32-page diagnostic algorithm and its one-page companion sit at the bottom of this page and remain the authoritative source.

1 HistoryOnset and course · domain-by-domain symptoms · collateral

Almost all of the diagnostic work is done here. Cognitive testing confirms and characterises what the history has already suggested — it rarely rescues a history that was never taken properly. Budget most of the consultation for this step.

Take a collateral history, separately where you can

An informant who knows the patient well is not optional — anosognosia is common, and in bvFTD and svPPA the patient's own account can be the least reliable part of the assessment. Interviewing patient and informant apart lets each speak freely; it also lets you compare accounts, which is itself diagnostic information.

  • What changed, and when did you first notice? Push for a specific anchoring event — a birthday, a holiday, a hospital admission — rather than accepting "a couple of years".
  • What can they no longer do that they could do before? Function is more informative than symptom lists, and it feeds directly into step 4.
  • Has it been steady, stepwise, fluctuating, or fast? The shape of the decline narrows the differential more than any single symptom.
  • Whose idea was this appointment? A patient dragged in by family, versus one who came worried on their own, points in different directions.

The shape of the decline

PatternTypical timeframeThink about
Insidious, slowly progressiveYearsAlzheimer's disease; FTD syndromes; PDD
Stepwise, with plateausAbrupt drops, then stabilityVascular cognitive impairment — look for the vascular events that match the steps
Fluctuating, hour to hour or day to dayMarked variability in attention and alertnessDLB (with the other core features); delirium (step 5) until proven otherwise
Rapid — weeks to months<12 months to significant disabilityRed flag (step 7): prion disease, autoimmune/limbic encephalitis, malignancy, rapidly progressive vascular disease
Static since a defined eventNo progressionPost-hypoxic, post-TBI, post-stroke — consider major NCD due to that cause, not a neurodegenerative process

Probe every domain, not just memory

Leading with memory questions biases the whole history towards Alzheimer's. Walk the six DSM-5-TR domains deliberately — SAMPLES on the Flowchart tab — asking for a concrete example of each.

  • Learning and memory — repeating questions, misplacing items, forgetting recent conversations or appointments. Ask whether reminders and cues help: benefit from cueing suggests a retrieval problem rather than a storage one.
  • Executive function — managing bills and the chequebook, following a recipe, planning a trip, coping when the routine changes, new poor judgement or scam vulnerability.
  • Complex attention — losing the thread in conversation, unable to do two things at once, needing quiet to function.
  • Language — word-finding pauses, circumlocution, substituting "thing", losing the meaning of words (svPPA), effortful halting speech with grammatical errors (nfvPPA).
  • Perceptual-motor — getting lost in familiar places, difficulty parking or judging gaps, trouble dressing, misreading or bumping into objects (consider posterior cortical atrophy).
  • Social cognition — loss of empathy or warmth, tactlessness, apathy mistaken for depression, new rigid routines, dietary change or sweet craving, disinhibition. Often the earliest change in bvFTD, and the one families report as "personality", not "memory".

Neuropsychiatric and sleep history

  • Visual hallucinations — well-formed, of people or animals, early in the course, points strongly to DLB. Distinguish from Charles Bonnet in visual loss.
  • REM sleep behaviour disorder — acting out dreams, shouting, falling out of bed, often years before cognitive change. Ask the bed partner; it can precede DLB or PD by a decade.
  • Depression, anxiety, apathy — carefully separate apathy (reduced initiation with intact mood) from depression, and consider whether mood is the cause of the cognitive picture rather than a consequence (step 6).
  • Delusions — theft, infidelity, misidentification, phantom boarder.
  • Sleep-disordered breathing — snoring, witnessed apnoeas, daytime somnolence; treatable and frequently missed.

Systems review that changes the differential

  • Gait and falls — early gait disturbance with urinary incontinence raises NPH; falls within the first year with parkinsonism raise PSP or DLB.
  • Continence — early urinary incontinence is atypical for AD and worth pursuing.
  • Autonomic — postural dizziness, syncope, constipation, erectile dysfunction (DLB, MSA).
  • Hearing and vision — uncorrected sensory loss both mimics and accelerates cognitive decline, and is one of the largest modifiable risk factors.
  • Head injury, seizures, stroke or TIA, cardiac and vascular history.

Medications, substances and risk background

  • Full medication reconciliation, including over-the-counter and complementary agents. Total anticholinergic burden, benzodiazepines, opioids, sedating antihistamines and antiemetics — note the start date against the timeline of decline.
  • Alcohol quantified honestly, plus nutritional risk (thiamine).
  • Vascular risk — hypertension, diabetes, AF, smoking, lipids, sleep apnoea.
  • Family history — ages at onset, and specifically any early-onset dementia, motor neurone disease, Huntington's, or a psychiatric-then-neurological pattern in a parent.
  • Education, literacy, occupation, first language and cultural background — you need these before you can interpret a single cognitive score in step 3.

Social, safety and legal

  • Living situation, who else is at home, and who is doing what for whom.
  • Carer identity, health and strain — the carer's capacity determines much of what is possible in step 10.
  • Driving, cooking and appliance safety, wandering, firearms access, and management of money — including anyone with access to accounts. See Fitness to drive.
  • Existing enduring power of attorney, advance health directive, or will — establish what already exists now, because capacity to make new ones is finite.
The Clinic companion runs this history, examination and formulation structure as an interactive checklist you can work through live during a consultation.
2 ExaminationLooking for the signs that move a diagnosis off Alzheimer's

In typical Alzheimer's disease the neurological examination is largely normal early on — which is exactly why it matters. Examination is mostly a search for the findings that say this is not typical AD, and for the treatable contributors.

General

  • Lying and standing blood pressure — postural drop supports DLB/MSA, and is a falls and delirium risk in its own right.
  • Weight and nutrition — unintentional loss is common in advancing dementia and is a prognostic marker; dietary change and sweet preference suggest bvFTD.
  • Hearing and vision, with the aids and glasses actually in use — test cognition only after correcting what you can. Cerumen impaction is worth excluding before you conclude anything.
  • Cardiovascular — irregular pulse (AF), carotid bruits, heart failure signs.
  • Signs of alcohol excess, chronic liver disease, thyroid disease.

Neurological signs and what they point to

FindingHow to elicitPoints towards
Focal deficit — asymmetric weakness, sensory loss, visual field defect, upgoing plantarStandard screening neurological examVascular disease, structural lesion — imaging becomes urgent (steps 7–8)
Parkinsonism — bradykinesia with rigidity and/or rest tremorFinger tapping, hand opening/closing, tone at wrist and elbow with contralateral activationDLB if within a year of cognitive onset; PDD if dementia follows established PD by >1 year; symmetrical and axial-predominant suggests PSP/MSA
Vertical supranuclear gaze palsy — downgaze firstPursuit then saccades in the vertical plane; check whether doll's-eye manoeuvre overcomes itProgressive supranuclear palsy — the single most useful bedside discriminator
Slow or hypometric saccades, apraxia of eyelid openingAsk for rapid gaze switches between two targetsPSP, corticobasal syndrome
Magnetic, wide-based, "stuck to the floor" gait with preserved arm swingObserve walking, turning, and tandem gaitNormal pressure hydrocephalus — especially with urinary incontinence
Shuffling, festinating gait with reduced arm swingAs aboveParkinsonian syndromes
Early falls, backward pull on retropulsionPull testPSP; DLB
Myoclonus, startle-sensitiveObserve at rest and on action; tap to startleRapidly progressive dementia — prion disease, but also late AD and metabolic causes
Limb apraxia, alien limb, cortical sensory loss, asymmetric dystoniaAsk to mime tool use; test graphaesthesia and stereognosisCorticobasal syndrome
Fasciculations, wasting, brisk reflexes with weaknessInspect tongue and limbs; full motor examFTD–motor neurone disease — changes prognosis and referral pathway immediately
Cerebellar signs — ataxia, nystagmus, dysarthriaHeel–shin, finger–nose, gaitMSA-C, alcohol, paraneoplastic, prion disease
Frontal release signs — grasp, palmomental, glabellar tapStandard elicitationNon-specific alone; supportive of frontal involvement when the rest of the picture fits
Peripheral neuropathyVibration, proprioception, ankle reflexesB12 deficiency, alcohol, diabetes — all worth chasing in step 6
Observation before formal examination earns its keep: how they walked in, whether they turned to the informant to answer (the head-turning sign), how they managed their coat and their handbag, whether they were oriented to the building. Much of the mental state examination is already done by the time you pick up the tendon hammer.
3 Cognitive assessmentChoosing the instrument · interpreting the profile, not just the total

A screening score is a measurement, not a diagnosis. Two patients can score 22 on the MoCA for entirely different reasons; which items they lost is worth more than the number they finished with.

Before you administer anything

  • Glasses and hearing aids on, quiet room, adequate light, no informant prompting from the corner.
  • Not in acute delirium, not acutely unwell, not in pain, not sedated — and note if any of these apply.
  • Test in the patient's first language wherever possible, with a professional interpreter rather than a family member. Translating a memory test on the fly invalidates it.
  • Record education, literacy, occupation and premorbid ability — you cannot interpret the result without them.

Choosing the instrument

Cut-offs are indicative screening thresholds, not diagnostic in isolation — interpret alongside age, education, language, sensory function, and premorbid ability.

ToolBest useTypical cut-off
MoCAPreferred first-line global screen in most guidelines; more sensitive than MMSE for MCI/mild NCD, executive & visuospatial deficits<26/30 (+1 pt if ≤12y education)
MMSEGlobal screen; insensitive to executive/frontal deficits; retained for PBS medication authority in Australia<24/30 · bands: 21–24 mild, 10–20 moderate, <10 severe
RUDASCulturally and linguistically diverse patients, low literacy — minimises language/education/cultural bias<23/30
KICA-Cog / KICA-ScreenUse in place of MMSE for Aboriginal and Torres Strait Islander patientsPer KICA scoring manual
Mini-Cog / Clock DrawingRapid bedside adjunct, not a standalone diagnostic instrumentAbnormal clock or 0/3 recall
ACE-IIISecond-line detailed domain breakdown; useful when subtype differentiation (e.g. AD vs. FTD profile) is needed<88/100 (indicative)
IQCODE-SFInformant-rated decline over 10 years; use when direct testing is unreliable or unavailableAverage item score >3.3–3.4 (scale 1–5; 3.0 = no change)

Reading the profile

  • Delayed recall that does not improve with cueing or recognition — a storage (hippocampal) deficit, characteristic of typical AD.
  • Recall that jumps with cueing — a retrieval deficit; think subcortical, vascular, frontal-executive, depression or medication effect.
  • Trails B, letter fluency, abstraction and serial subtraction disproportionately affected — executive/frontal-subcortical pattern.
  • Naming and single-word comprehension lost with fluent, empty speech — semantic.
  • Cube, clock and intersecting pentagons disproportionately affected — visuospatial; consider DLB or posterior cortical atrophy.
  • Effortful, agrammatic speech with speech apraxia — nfvPPA.
  • A near-normal score with a family reporting profound change — take the family seriously. Screening instruments are weakest at exactly the two ends where you most need them: high premorbid ability, and bvFTD, where MMSE and MoCA can both be normal.

When to go further

  • Formal neuropsychological assessment — young onset, high premorbid ability with normal screening but convincing history, ambiguous mild NCD, suspected mimic (particularly depression), medicolegal or capacity questions, or when subtype differentiation will change management.
  • Speech pathology assessment — suspected primary progressive aphasia, where the variant classification often turns on language features a screening tool does not sample.
  • Repeat testing — when the diagnosis is uncertain, a documented change over 6–12 months is often more informative than another cross-sectional test today. Use the same instrument, and mind practice effects.

Pattern-recognition tools

Four bedside references, each one opening in place: the screening instruments and how to read their profiles, a heat-map of which features carry weight in which condition, and criteria-oriented checklists for the individual subtypes. Compiled by Rhymie Yusoff from the sources on the Education page.

Dementia cognitive tests & subtype interpretation patterns

How to use it

1Start broad

MoCA, MMSE, RUDAS, Mini-Cog or ACE-III, chosen for the setting, language, education and the time you have.

2Localise the deficit

Add focused tasks for memory storage, executive control, language, visuospatial function, praxis or social cognition.

3Read the pattern

Interpretation rests less on the total than on which domains are disproportionately impaired, how fast function changed, and whether fluctuation, parkinsonism, behavioural change or focal signs sit alongside.

Clinical principles

No screening tool diagnoses dementia on its own. A result has to be read against collateral history, a delirium screen, mood, medication burden, sensory status, education and language background, the neurological examination, functional decline, and imaging or biomarkers where relevant.

What scores mean
  • Totals track severity; domain-level errors usually carry more diagnostic weight than the cut-off.
  • A low score without functional decline may be delirium, depression, low education, aphasia, anxiety or a sensory barrier — not neurodegeneration.
  • Serial decline matters more than one mildly abnormal result.
Pattern recognition
  • Amnestic storage failure points to the Alzheimer spectrum.
  • Executive slowing with patchy retrieval but preserved recognition suggests vascular or subcortical disease.
  • Visuospatial and attentional deficits out of proportion to memory raise Lewy body disease or a posterior cortical syndrome.
Common traps
  • Don't overcall dementia during acute illness or after sedation.
  • Don't compare MMSE and MoCA scores as if interchangeable.
  • Language-led and behavioural syndromes can score deceptively well on memory-heavy screeners early.

Core screening tests

Choose the instrument that matches the patient's baseline language, literacy, cultural context, stamina and the profile you suspect.

MoCA 10 to 15 min Sensitive to early change Executive heavy

The Montreal Cognitive Assessment is often preferred for mild cognitive impairment and early dementia because it samples executive and visuospatial tasks more broadly than the MMSE.

  • Key domains: attention, executive function, visuospatial skills, naming, memory learning and delayed recall, language, abstraction, orientation.
  • Particularly useful when frontal-subcortical or Lewy body patterns are suspected.
  • Education correction is commonly applied, but interpretation should still be contextual rather than formulaic.
Typical broad cut point
< 26/30
Pattern clue
Poor delayed recall with weak cueing suggests storage failure
Limitation
Language, vision, tremor, and education can affect performance
MMSE 7 to 10 min Widely recognised Less frontal sensitivity

The Mini-Mental State Examination remains familiar and practical for longitudinal monitoring, although it is less sensitive to executive dysfunction and mild impairment.

  • Strongest for orientation, registration, recall, language, and simple visuoconstruction.
  • May underestimate deficits in early frontotemporal, vascular, or Lewy body presentations.
  • Useful where historical comparability matters across services.
Typical broad cut point
< 24/30
Pattern clue
Disproportionate orientation and recall loss fits Alzheimer spectrum disease
Limitation
Ceiling effect in highly educated or very early disease
ACE-III 15 to 20 min Profile rich Helpful in atypical syndromes

The Addenbrooke’s Cognitive Examination offers a broader profile than short screeners and is especially useful when language, semantic knowledge, or frontal features may be important.

  • Domains: attention, memory, fluency, language, visuospatial function.
  • Fluency and language subscales can be informative in frontotemporal and primary progressive aphasia syndromes.
  • Better suited to clinic than very acute bedside settings when time is available.
Broad cut points often used
< 82 or < 88/100
Pattern clue
Fluency-language drop may point to frontal or temporal degeneration
Limitation
Needs more time and patient stamina
RUDAS 10 min Cross-cultural utility Less education bias

The Rowland Universal Dementia Assessment Scale is valuable when cultural and linguistic fairness is a priority and can be preferable to MMSE or MoCA in multilingual populations.

  • Assesses memory, praxis, body orientation, judgement, language, and visuoconstruction.
  • Useful in hospital and community geriatrics where English is not the patient’s first language.
  • Still requires interpreter support and thoughtful administration where needed.
Typical broad cut point
< 23/30
Pattern clue
Good alternative when low scores on other tools may reflect language unfairness
Limitation
Less detailed subtype profiling than ACE-III
Mini-Cog 3 min Very brief Screen only

The Mini-Cog is a fast case-finding tool rather than a detailed cognitive profile. It is useful in busy acute, perioperative, or primary care settings.

  • Combines three-word recall with clock drawing.
  • Helpful when you need a rapid first-pass screen before fuller assessment.
  • Abnormal results should trigger a more detailed tool, not be treated as a subtype discriminator.
Typical interpretation
0 to 2 usually abnormal; 3 to 5 lower concern depending on recall and clock
Pattern clue
Clock failure may flag executive or visuospatial dysfunction
Limitation
Too coarse for nuanced dementia typing
Clock Drawing Test 2 to 3 min Visual-executive clue Useful adjunct

The clock test is not a stand-alone diagnostic instrument but often provides a quick window into executive planning, visuospatial construction, comprehension, and attention.

  • Look beyond pass or fail: number placement, spatial crowding, perseveration, omission, and hand setting all matter.
  • Particularly informative in Lewy body disease, vascular cognitive impairment, and posterior cortical dysfunction.
  • Performance is heavily influenced by vision, hemineglect, motor impairment, and comprehension.
Interpretation style
Qualitative errors often more useful than numeric score
Pattern clue
Spatial disorganisation suggests parietal or visuospatial dysfunction
Limitation
Not specific for dementia subtype on its own

Focused domain tests

Once the screen suggests a direction, these decide whether the problem is storage, retrieval, executive control, language, praxis, social cognition or visuospatial processing.

Memory-focused
  • List learning (Hopkins Verbal Learning Test or similar) — poor delayed recall plus poor recognition supports encoding or storage failure.
  • Free and cued selective reminding — failure to improve with cueing suggests hippocampal storage loss rather than retrieval failure.
  • Story recall shows amnestic decline but is influenced by language load.
Executive-focused
  • Trail Making A and B — slowed set shifting points to frontal-subcortical dysfunction.
  • Digit span backward, months backward, letter fluency — bedside detection of attention and executive inefficiency.
  • Stroop-type interference or go/no-go tasks are especially useful in suspected bvFTD.
Language-focused
  • Category versus letter fluency dissociation helps separate semantic from frontal syndromes.
  • Boston Naming Test or a brief naming task detects anomia.
  • Sentence repetition, single-word comprehension and grammar testing are central when PPA is suspected.
Visuospatial and praxis
  • Figure copy, intersecting pentagons, Rey complex figure and clock drawing expose spatial and constructional deficits.
  • Limb and buccofacial praxis testing supports corticobasal or parietal syndromes when apraxia is prominent.
  • Visual object and face recognition help when posterior cortical atrophy or semantic dementia is in the differential.

Pattern interpretation matrix

A heuristic, not a rulebook. Mixed pathology is common in older patients, and late disease blurs the classic distinctions.

SubtypeScreening patternHigh-yield extra cluesInterpretive pitfall
Alzheimer disease
Typical amnestic presentation
Poor delayed recall out of proportion to other domains; cueing and recognition often remain impaired; orientation declines as disease progresses.Storage failure, repetition, topographical disorientation, later language and visuospatial involvement.May look mild on very brief testing early, especially in highly educated patients.
Vascular cognitive impairmentExecutive dysfunction, psychomotor slowing, impaired attention, variable recall with better recognition than free recall.Focal deficits, gait disorder, urinary symptoms, stepwise decline, pyramidal signs, imaging burden.Can be mistaken for depression, frailty, or nonspecific slowing if executive testing is weak.
Dementia with Lewy bodiesVisuospatial, attentional, and executive deficits often exceed memory loss early; clock and figure copy may be poor.Cognitive fluctuation, recurrent visual hallucinations, REM sleep behaviour disorder, parkinsonism, neuroleptic sensitivity.Memory may be near-normal early, which can falsely reassure clinicians using memory-dominant screens.
Parkinson disease dementiaFrontal-subcortical pattern: slowed processing, set-shifting difficulty, reduced fluency, visuospatial inefficiency, later memory retrieval issues.Established Parkinson disease preceding dementia by more than a year.Motor disability may depress test performance independent of cognition.
Behavioural variant FTDExecutive and fluency impairment may appear with relatively preserved orientation and memory early; social cognition deficits are often missed on standard screeners.Disinhibition, apathy, loss of empathy, compulsions, dietary change, poor insight.Short cognitive tests can be near normal despite major behavioural and functional decline.
Primary progressive aphasia
language-led FTD spectrum
Language tasks disproportionately abnormal; repetition, naming, grammar, or word meaning fail depending on subtype.Speech production change, word-finding pauses, comprehension loss, spared memory for nonverbal material early.Low totals may reflect aphasia load more than global dementia severity.
Posterior cortical atrophyMarked visuospatial and visuoperceptual failure; drawing, copying, line orientation, and visual search are poor.Simultanagnosia, alexia, dressing difficulty, visual crowding, preserved insight early.Can be mislabelled as eye disease or anxiety before cognitive assessment is expanded.
Corticobasal syndrome / PSP-related syndromesExecutive dysfunction, slowed processing, praxis failure, visuospatial deficits; may have language or behavioural overlay.Asymmetry, apraxia, dystonia, gaze palsy, falls, frontal behavioural change.Motor and ocular deficits can contaminate paper-based tasks.

Subtype-specific bedside reading

Alzheimer disease

The hallmark is episodic memory storage failure. Patients learn poorly, forget rapidly, and often do not improve much with cueing or recognition prompts.

  • Best clues: delayed recall, orientation, later naming and visuospatial decline.
  • Helpful tools: MoCA, ACE-III, list learning with recognition.
  • Interpretation pattern: amnestic profile outweighs executive inefficiency early.
Vascular cognitive impairment

The signature is frontal-subcortical inefficiency rather than pure storage loss. Retrieval is poor, but recognition is often relatively better than spontaneous recall.

  • Best clues: slowed Trails, reduced fluency, attention lapses, dysexecutive clock drawing.
  • Helpful tools: MoCA, Trails, verbal fluency, digit span, gait plus neuro exam.
  • Interpretation pattern: patchy deficits with variable day-to-day performance.
Dementia with Lewy bodies

Attention, executive function, and visuospatial ability are often disproportionately impaired, while verbal memory may be less affected early than expected for the level of disability.

  • Best clues: fluctuating attention, poor clock or cube copy, hallucinations, RBD history.
  • Helpful tools: MoCA, clock drawing, figure copy, attentional tasks repeated over time.
  • Interpretation pattern: think of DLB when function is worse than the memory score suggests.
Behavioural variant FTD

Behaviour changes often precede obvious deficits on generic screeners. Executive control, abstraction, fluency, and social cognition are the most informative domains.

  • Best clues: poor verbal fluency, rule breaking, impulsive responses, preserved orientation.
  • Helpful tools: ACE-III, frontal batteries, fluency, social cognition probes, collateral history.
  • Interpretation pattern: normalish MMSE does not exclude major frontal degeneration.
Primary progressive aphasia subtypes

The pattern depends on whether the dominant problem is agrammatism and effortful speech, impaired repetition, or semantic loss.

  • Non-fluent variant: grammar and speech production fail; repetition may be impaired.
  • Logopenic variant: word-finding pauses and sentence repetition difficulty are prominent; often Alzheimer pathology.
  • Semantic variant: naming and single-word comprehension deteriorate with surface dyslexia and object knowledge loss.
Parkinson disease dementia

The profile usually resembles other subcortical syndromes: bradyphrenia, executive dysfunction, visuospatial weakness, and retrieval-based memory problems.

  • Best clues: processing speed, fluency, attentional set shifting, visuospatial construction.
  • Helpful tools: MoCA, Trails, clock drawing, dual-task observation, collateral timeline.
  • Interpretation pattern: use the one-year rule to separate it from DLB in syndromic classification.

Red flags for a non-degenerative or mixed cause

  • Acute or fluctuating onset, reduced arousal, or inattention — suggesting delirium.
  • Prominent low mood, psychomotor retardation or inconsistent effort — suggesting depression or a functional cognitive disorder.
  • Marked asymmetry, upper motor neurone signs, aphasia out of keeping with the rest of the profile, or rapid decline — suggesting a structural, inflammatory or prion-like process.

Practical reporting language

  • Describe domains, not only totals: "performance shows predominant amnestic storage impairment with secondary executive inefficiency."
  • State limitations explicitly: hearing loss, interpreter use, severe tremor, fatigue, pain, visual impairment.
  • Separate syndrome from aetiology: "major neurocognitive disorder with a dysexecutive-visuospatial pattern, suspicious for Lewy body disease."
Best used alongside collateral history, functional assessment and the neurological examination — see Clinic companion.
Dementia hallmark heat-map — feature by condition
– Not typical + Can occur ++ Common +++ Hallmark
Feature AD VaD DLB bvFTD PPA PCA PDD CJD HIV Dialysis NPH Mixed
Memory
Early episodic memory lossGradual +++ + + + + + + + + ++
Executive / attention
Executive dysfunction earlyVariable + +++ ++ +++ + + +++ ++ +++ ++ ++ ++
Slowed processing speedVariable + +++ ++ + + +++ ++ +++ ++ ++ ++
Stepwise or stroke-linked declineStepwise +++ ++
Visuospatial or visual-posterior dysfunctionVariable + + +++ +++ ++ + + + +
Neuropsychiatric
Cognitive fluctuationsFluctuating + +++ ++ + + + +
Visual hallucinationsFluctuating + +++ + ++ +
Motor / sleep
REM sleep behaviour disorderOften precedes +++ ++
ParkinsonismProgressive + +++ + +++ ++ + + + +
Myoclonus or startle phenomenaRapid +++ ++
Ataxia or cerebellar involvementRapid / variable + +++ +
Language / behaviour
Behavioural disinhibition or apathyProgressive + + + +++ + + + + + + +
Language-predominant syndromeProgressive + + ++ +++ + +
Word comprehension or semantic lossProgressive + +++
Effortful nonfluent speech or agrammatismProgressive + +++
Gait / autonomic
Early gait disturbanceVariable +++ ++ + + +++ ++ ++ ++ +++ ++
Early urinary symptomsVariable ++ + + +++ ++
Relative early memory preservationRelative + + ++ +++ +++ + + +
The columns
AD
Alzheimer diseaseSlow, gradual — 8–10y from diagnosis
VaD
Vascular dementiaVariable, classically stepwise
DLB
Dementia with Lewy bodiesModerate — about 5–8y
bvFTD
Behavioural-variant FTDModerate to fast — 6–9y; 2–3y with MND overlap
PPA
Primary progressive aphasiaVariable — 7–10y; logopenic tracks like AD
PCA
Posterior cortical atrophyGradual; usually younger onset than typical AD
PDD
Parkinson disease dementiaModerate — marks more advanced PD
CJD
Creutzfeldt–Jakob diseaseVery fast — median 4–6 months, sporadic
HIV
HIV-associated neurocognitive disorderVariable; may stabilise or improve on treatment
Dialysis
Dialysis-associated encephalopathySubacute; related to dialysis and metabolic state
NPH
Normal pressure hydrocephalusSubacute — potentially shunt-responsive
Mixed
Mixed dementia (AD + vascular)Moderate — often faster than pure AD
Hover or tap any cell for why that feature carries the weight it does in that condition. A hallmark is a strong clue, never a diagnosis on its own: read a row to see how far a single finding narrows the field, and a column to see the shape of a syndrome. Mixed pathology is common in older patients and late disease blurs all of these distinctions.
Dementia subtype interactive toolkit

Eleven criteria-oriented checklists. Tick what applies and each one reads back what the pattern supports. These are recognition aids built from the published criteria — they are not the criteria themselves, and nothing typed here is stored or sent anywhere.

Alzheimer-pattern diagnostic pathway

Pattern checklist; biomarkers determine biological confirmation.

0 of 7 selected

Dementia with Lewy bodies

Consensus diagnostic-criteria calculator; use the DIAMOND-Lewy toolkit for structured clinical questioning.

0 of 8 selected

Behavioural-variant frontotemporal dementia

Rascovsky/FTDC-oriented screen. The first six items are core behavioural/cognitive features.

0 of 8 selected

Primary progressive aphasia and variants

First establish that language decline is progressive, prominent and the main cause of impaired daily activity.

0 of 10 selected

Vascular cognitive disorder

VASCOG-oriented screen; neuroimaging and clinicoradiological correlation are essential.

0 of 8 selected

Parkinson disease dementia versus DLB

The one-year timing convention separates the clinical labels; both are Lewy-body diseases.

0 of 6 selected

Progressive supranuclear palsy

MDS-PSP-oriented recognition checklist; formal criteria use graded certainty and phenotype combinations.

0 of 7 selected

Corticobasal syndrome

Clinical syndrome checklist; corticobasal syndrome has several possible underlying pathologies.

0 of 7 selected

Idiopathic normal-pressure hydrocephalus

Recognition and referral checklist; diagnosis requires imaging and specialist assessment.

0 of 6 selected

Huntington-related neurocognitive disorder

Use only in the appropriate neurological and genetic context.

0 of 6 selected

Rapidly progressive dementia / CJD red flags

Urgent pathway—not a routine dementia subtype score.

0 of 8 selected

Core references

Work from current local guidelines and the original publications before making or documenting a diagnosis. A checklist can support a pattern; it cannot establish one.
Cognitive pattern → subtype glossary

Plain-language terms used across the pattern-recognition tools above, from the companion selector tool.

Encoding / storage / retrieval / cueing
Encoding is taking in new information; storage is keeping it; retrieval is bringing it back to mind. Improvement with cues points to a retrieval, not storage, problem.
Episodic vs. semantic memory
Episodic memory is for personally experienced events tied to time and place; semantic memory is knowledge of what words and concepts mean.
Executive function / set-shifting
Mental control used to plan, organise, switch strategies, and monitor mistakes; set-shifting is flexibly changing from one rule or task to another.
Agrammatism / apraxia of speech / dysarthria
Agrammatism is difficulty with grammar; apraxia of speech is difficulty planning speech movements; dysarthria is unclear speech from weak or poorly coordinated muscles — not a language problem.
Praxis / apraxia / agnosia
Praxis is performing learned purposeful actions; apraxia is difficulty doing so despite adequate strength; agnosia is difficulty recognising a familiar object, face, or sound despite adequate sensation.
Disinhibition / apathy
Disinhibition is reduced control over impulsive or socially inappropriate behaviour; apathy is reduced motivation or interest, not automatically depression.
Mixed pathology
More than one disease process contributing to cognitive decline — for example, Alzheimer and vascular changes occurring together.
Pattern recognition is not diagnosis. Acute or fluctuating confusion, reduced alertness, or new focal neurological symptoms requires urgent assessment.
4 Functional assessmentThe step that decides major versus mild NCD

DSM-5-TR Criterion B turns entirely on function: does the cognitive deficit interfere with independence in everyday activities? If yes, it is major NCD; if the person compensates but is still independent, it is mild. This is a clinical judgement about a specific person's real life — no test score makes it for you.

Instrumental ADLs go first

Complex, cognitively demanding tasks fail before basic self-care does. Ask about each one concretely, and ask the informant, not just the patient.

  • Finances — paying bills on time, banking, unusual purchases, susceptibility to scams, unopened mail. Often the very first IADL to go, and the highest-consequence.
  • Medications — missed or doubled doses, whether a blister pack has become necessary, whether someone else now manages them.
  • Transport — driving incidents, getting lost, giving up night driving, stopping altogether.
  • Shopping and cooking — repeat purchases, spoiled food in the fridge, meals simplified or skipped, appliances left on.
  • Telephone and technology — managing calls, messages, and any device they previously used confidently.
  • Housework and home maintenance — a decline in standards that the person has not noticed.

Then basic ADLs

  • Bathing, dressing, grooming, toileting, transfers, feeding, continence — typically preserved until moderate disease, so early loss here should prompt a search for a physical cause or a non-AD diagnosis.

Instruments worth using

InstrumentCoversNotes
Lawton–Brody IADL8 instrumental activitiesThe workhorse for the major/mild distinction; interpret the historical gender-scored items with care
Katz ADL / Barthel Index6–10 basic self-care itemsBetter suited to moderate–severe disease and to care planning than to diagnosis
FAQ (Functional Activities Questionnaire)10 informant-rated complex activitiesSensitive to the mild NCD to major NCD transition; widely used in research cohorts
Bristol ADL Scale20 items, carer-rated, dementia-specificGood for tracking change over time and for identifying targets for occupational therapy
Occupational therapy functional assessmentObserved performance, ideally in the homeThe most informative option where the history is contested or the patient minimises — and it generates the intervention plan at the same time

Before you attribute a functional loss to cognition

  • Is it physical? Arthritis, weakness, breathlessness, pain and frailty all cause functional loss without cognitive impairment.
  • Is it sensory? Uncorrected hearing or vision loss disables independently.
  • Is it motivational or mood-related? Apathy and depression reduce performance without reducing capability.
  • Is it environmental or lifelong? Someone whose spouse always did the banking has not lost a skill they never exercised — ask what changed, not just what they cannot do.

Turning the assessment into a plan

  • Refer for My Aged Care assessment for home supports and, where appropriate, a comprehensive assessment for higher-level care — the wait times mean starting early is part of the treatment.
  • Occupational therapy home assessment for safety, equipment and simplification of high-risk tasks.
  • Name explicitly which functions are now unsafe to perform unsupervised, and document it — particularly driving, cooking, medication management and finances.
  • Re-measure function at review; functional trajectory is the most meaningful progression marker you have, and more useful than a repeated screening score.
5 Delirium exclusionDSM Criterion C — do not stage or subtype through a delirium

Neither severity nor aetiology can be finalised until delirium has been excluded or has resolved. A cognitive score obtained during a delirium tells you about the delirium, not about the underlying brain — and a diagnosis of dementia made on that score will follow the patient for the rest of their life.

Screen everyone at first contact

ToolItems scoredPositive result
4ATAlertness; AMT4 (age, DOB, place, year); Attention (months backwards); acute change/fluctuation over 2 weeks≥4 = probable delirium; 1–3 = possible cognitive impairment; 0 = delirium unlikely (but doesn't fully exclude it if attention couldn't be tested)
CAMFeature 1: acute onset/fluctuating course; Feature 2: inattention; Feature 3: disorganised thinking; Feature 4: altered consciousnessPositive = Feature 1 AND 2, plus either 3 or 4

What separates delirium from dementia at the bedside

DeliriumDementia
OnsetHours to daysMonths to years
CourseFluctuates markedly, often worse at nightSlowly progressive; day-to-day variation is modest — except in DLB
AttentionImpaired — the cardinal featureRelatively preserved until later stages
Consciousness / alertnessAltered — hyperalert or drowsyNormal until advanced
ReversibilityUsually reversible if the cause is found and treatedNot reversible, though contributors can be
Hypoactive delirium is the one that gets missed. A quiet, withdrawn, drowsy patient who is "just settled" or "just depressed" is the commonest presentation in older inpatients and carries the worse prognosis. Fluctuating inattention with reduced alertness is delirium until you have proved otherwise.

Delirium superimposed on dementia

  • Dementia is the single largest risk factor for delirium, and the two coexist constantly — finding one never excludes the other.
  • The diagnostic question is not "which is it?" but "what is the baseline, and what is new?" Establishing the premorbid baseline from an informant, using something like the IQCODE, is what makes the distinction possible.
  • Where a delirium is present, treat it, then re-assess cognition once it has resolved — commonly weeks to months later, not at discharge.
  • An episode of delirium is itself a marker of substantially increased risk of subsequent dementia diagnosis; arrange follow-up cognitive review even when the patient returns to apparent baseline.

Look for the precipitant

The PITCHED checklist in step 10 works just as well as a delirium precipitant screen as it does for behavioural symptoms — pain, infection, thirst, constipation, hunger, environment and drugs account for a large share of what precipitates delirium in this population. Add hypoxia, retention, electrolyte disturbance, and any recent medication change.

6 Mimics and contributorsDSM Criterion D — and the treatable things you do not want to miss

The decline should not be better explained by another mental disorder or by a reversible cause. Truly reversible dementias are uncommon, but partially reversible contributors sit on top of a neurodegenerative process very often — and treating them is frequently the largest single improvement you can offer.

The standard screen

  • Bloods: FBC, EUC, LFTs, calcium, glucose or HbA1c, TSH, vitamin B12 and folate.
  • Consider by indication: ESR/CRP, syphilis and HIV serology, vitamin D, coeliac serology, copper/caeruloplasmin in young onset, drug levels, and a toxicology screen where substance use is plausible.
  • Structural neuroimaging — step 8.
  • Sleep study where sleep-disordered breathing is suspected.

Mimics to screen for

  • Depression / pseudodementia
    Often subacute onset, prominent "I don't know" answers, effort-dependent testing, mood symptoms preceding cognitive complaints — improves with treatment of the mood disorder. Late-life depression can also be a prodrome of dementia, so treat the mood and re-test rather than choosing between the two.
  • Other psychiatric disorders
    Anxiety, psychotic disorders, and severe distress can all impair attention and processing speed on bedside testing.
  • Substance or medication effects
    Anticholinergics, benzodiazepines, opioids, alcohol, sedating antihistamines and antiepileptics. Check the temporal relationship against the decline, calculate the anticholinergic burden, and run a structured deprescribing trial before accepting the diagnosis.
  • Uncorrected sensory impairment
    Hearing or visual loss can masquerade as cognitive impairment on bedside testing, and independently accelerates decline. Correct it, then re-test.
  • Delirium not yet resolved
    Revisit step 5 if the picture is fluctuating or acute.
  • Endocrine and metabolic
    Hypothyroidism, hypercalcaemia, hyponatraemia, hypoglycaemia and B12 deficiency — all on the standard screen, all worth correcting even when they are clearly not the whole story.
  • Normal pressure hydrocephalus
    Gait disturbance first, then urinary incontinence, then cognition. Ventriculomegaly out of proportion to sulcal atrophy with a tight high convexity (DESH) on imaging — refer for tap test and neurosurgical opinion.
  • Chronic subdural haematoma
    Often no recalled head injury, particularly on anticoagulants. Fluctuating conscious state or focal signs; found on the scan you ordered in step 8.
  • Obstructive sleep apnoea
    Impairs attention, memory consolidation and executive function; treatment produces measurable cognitive gains.
  • Alcohol-related brain injury and thiamine deficiency
    Consider Wernicke encephalopathy in any malnourished or alcohol-dependent patient with confusion, ataxia or eye movement abnormality — treat with parenteral thiamine first and confirm later.
  • Chronic pain and its treatment
    Both the pain and the opioid can be doing the damage; each is modifiable.
A useful discipline: before you write the diagnosis, list every contributor you have found and state what you are doing about each. Hearing aids, a deprescribing plan, treating the depression and fixing the sleep apnoea will not reverse Alzheimer's disease — but they routinely change what the person can do.
7 Red flagsFeatures that change the workup, the urgency, or the destination

Most of dementia assessment is unhurried and outpatient. These are the features that should pull a patient out of that pathway — because the diagnosis is likely to be something other than a common neurodegenerative dementia, because the investigation set is different, or because the window for treatment is short.

Red flagWhat it raisesWhat to do
Onset before 65Genetic and inherited causes, FTD spectrum, autoimmune, metabolic, HIV, alcohol-related, Huntington's, Wilson'sRefer to a young-onset dementia or cognitive neurology service; MRI rather than CT; broaden the screen; take a three-generation family history and consider genetic counselling
Rapid progression — significant decline over weeks to monthsPrion disease, autoimmune/limbic encephalitis, CNS malignancy or lymphoma, vasculitis, rapidly progressive vascular disease, metabolicExpedited or inpatient workup: urgent MRI with DWI, EEG, lumbar puncture, autoimmune and paraneoplastic antibody panels, malignancy screen. Neurology involvement early
Early prominent focal neurological signs — hemiparesis, visual field loss, aphasia out of keeping with the cognitive pictureStructural lesion, stroke, tumour, subduralUrgent neuroimaging before further cognitive workup
New seizuresStructural lesion, autoimmune encephalitis (especially faciobrachial dystonic seizures with LGI1), late-onset epilepsy in ADNeuroimaging, EEG, neurology referral; faciobrachial dystonic seizures warrant urgent autoimmune workup
Myoclonus or startle myoclonus, earlyPrion disease; also late AD, metabolic and drug causesMRI with DWI looking for cortical ribboning and basal ganglia change; EEG; CSF including RT-QuIC via a specialist service
Prominent early gait disorder with urinary incontinenceNormal pressure hydrocephalus; also vascular, PSPImaging looking for DESH; neurosurgical referral for tap test if the pattern fits
Early falls with parkinsonism, or vertical gaze palsyPSP, corticobasal syndrome, MSAMovement disorder or neurology referral; imaging for midbrain atrophy
Early autonomic failure — marked postural hypotension, syncope, urinary or erectile dysfunctionMSA, DLBAutonomic assessment; avoid antipsychotics until DLB is excluded
Behavioural change with pyramidal signs, fasciculations or bulbar symptomsFTD–motor neurone diseaseUrgent neurology referral — prognosis and the entire care plan change
Headache, papilloedema, vomiting, or progressive focal signsRaised intracranial pressure, tumourUrgent imaging
Immunosuppression, known malignancy, or recent immunotherapyOpportunistic infection, CNS metastases, paraneoplastic or immune-related encephalitisUrgent imaging, CSF, and discussion with the treating team
Fever, meningism, or systemic inflammatory featuresCNS infection, vasculitisAcute pathway — do not manage as outpatient dementia
Abrupt psychiatric onset in a previously well older adult with rapid cognitive changeAutoimmune encephalitis, paraneoplastic syndromesAntibody panels, MRI, EEG, CSF; treat early — delay costs recovery
Anticoagulation with fluctuating cognition or conscious stateChronic subdural haematomaNon-contrast CT head, low threshold
Family history of early-onset dementia, MND, Huntington's or prion diseaseAutosomal dominant causeGenetic counselling before any testing; specialist service referral
The imaging counterparts of these — what NPH, CJD, limbic encephalitis, subdural, Wernicke and tumour actually look like on the scan, with the sequences and the signs to ask for — are in Clinical anatomy and neuroradiology of dementia, under Imaging red flags.
8 NeuroimagingWhat to request, what to ask the radiologist for, what to read yourself

Structural imaging is part of the standard workup for a new diagnosis of dementia, for two separate reasons: to exclude a lesion that changes management, and to support or refute the suspected subtype. It does not make the diagnosis — a normal scan does not exclude dementia, and atrophy on a scan does not create it.

Which modality

  • MRI is preferred where it is available and tolerated — better grey/white differentiation, better assessment of the medial temporal lobes, and the only way to see microbleeds, restricted diffusion and small infarcts properly.
  • CT is acceptable when MRI is contraindicated, unavailable, or not tolerated. It will still exclude a tumour, subdural or hydrocephalus, and gives a usable impression of atrophy and vascular burden.
  • Where the clinical question is urgent (step 7), image on the urgent pathway rather than waiting for an outpatient MRI slot.

What to ask for on the request

  • Coronal T1 perpendicular to the long axis of the hippocampus — the plane in which medial temporal atrophy is actually scored. Ask for it by name; you will not get it by default.
  • FLAIR for white matter hyperintensity burden.
  • T2*-weighted or SWI for microbleeds and superficial siderosis — essential if cerebral amyloid angiopathy is a consideration, and mandatory before any anti-amyloid therapy.
  • DWI where rapid progression or prion disease is in the differential.
  • State the clinical question on the request. "Cognitive decline" gets you a report about atrophy; "suspected svPPA — assess anterior temporal asymmetry" gets you an answer.

Read it yourself, with the rating scales

Four visual rating scales cover most of what a structural scan contributes in this population. All four, with real template sections and worked grades, are in the neuroradiology pane.

ScalePlaneWhat it tells you
MTA (Scheltens, 0–4)Coronal T1 at the hippocampal bodyMedial temporal atrophy — supports AD; age-adjusted thresholds matter (≥2 abnormal under 75, ≥3 over 75)
PA (Koedam, 0–3)Sagittal, axial and coronal togetherPosterior/parietal atrophy — younger-onset AD and posterior cortical atrophy, where MTA can be normal
GCA and GCA-F (0–3)AxialGlobal cortical atrophy and its frontal subscale — frontal predominance supports bvFTD
Fazekas (0–3)Axial FLAIRWhite matter hyperintensity burden — small vessel disease contribution

Patterns worth knowing

  • Alzheimer's — medial temporal and posterior/parietal atrophy, usually symmetric.
  • bvFTD — frontal and anterior temporal atrophy, often strikingly asymmetric.
  • svPPA — asymmetric anterior temporal lobe atrophy, typically left-predominant, with a "knife-edge" appearance.
  • Vascular — strategic infarcts, lacunes, confluent white matter change; the burden should plausibly explain the deficit.
  • DLB — relatively preserved medial temporal lobes compared with AD of similar severity.
  • PSP — midbrain atrophy, the hummingbird and morning-glory signs.
  • CAA — lobar microbleeds and cortical superficial siderosis on SWI.

Beyond structural imaging

InvestigationUseAustralian access
FDG-PETHypometabolic pattern where structural imaging and clinical picture disagree — particularly AD versus FTDMedicare rebate available under specific criteria; specialist referral
Amyloid PETConfirms or excludes amyloid pathology; now a gateway investigation for anti-amyloid therapyNot generally Medicare-rebated for routine diagnosis; largely self-funded or trial-based
DaT-SPECTReduced striatal dopamine transporter uptake — an indicative biomarker for DLB, helps separate it from ADAvailable in major centres; limited rebate
MIBG cardiac scintigraphyReduced uptake in DLB — an alternative indicative biomarkerLimited availability
CSF biomarkers (Aβ42/40, p-tau, total tau)Where the diagnosis remains genuinely uncertain, or before anti-amyloid therapySpecialist services; not routine
Blood-based biomarkers (plasma p-tau217)Rapidly maturing; strong performance for amyloid status in specialist settingsEmerging — treat results outside a specialist pathway with caution, and do not use to diagnose in isolation
PolysomnographyConfirms REM sleep behaviour disorder (supportive of DLB) and identifies sleep-disordered breathingWidely available
The full neuroradiology material — an interactive 3D model of the regions each subtype involves, real T1 sections through the standard planes, the four visual rating scales with worked grades, and the imaging red flags — is in Clinical anatomy and neuroradiology of dementia.
9 Formulation and DSM-5-TR diagnosisMajor or mild · severity · aetiological subtype · behavioural specifier

Everything above now assembles into a diagnosis with four parts: whether it is a neurocognitive disorder, whether it is major or mild, how severe it is, and what is causing it.

The four criteria, in summary

CriterionWhere you established it
AEvidence of decline in one or more of the six cognitive domains, from concern plus objective testingSteps 1 and 3
BDoes the deficit interfere with independence in everyday activities? — yes for major, no for mildStep 4
CNot occurring exclusively in the context of deliriumStep 5
DNot better explained by another mental disorderStep 6

The Flowchart tab works through each criterion in full, including the SAMPLES domain list and the major-versus-mild distinction.

Severity — and staging with CDR

DSM-5-TR specifies severity for major NCD as mild, moderate or severe, defined by which activities need assistance: instrumental activities (mild), basic activities (moderate), or full dependence (severe). The Clinical Dementia Rating is a clinician-rated global stage that maps onto this and is more reproducible for tracking over time — use it to stage once the diagnosis is made, not to make it.

Global CDRStageTypical CDR-SB range
0Normal0
0.5Questionable / very mild (MCI / Mild NCD range)0.5–4
1Mild dementia (Major NCD)4.5–9
2Moderate dementia9.5–15.5
3Severe dementia16–18
CDR and driving risk — what the evidence does and does not support
Global CDRRelative risk of failing an on-road driving test, vs. CDR 0
0.5RR 82.7 in one Class I study; RR 9.67–25 in others
1RR 88.67 (up to 12 in other studies); also more likely judged unsafe on 6-month follow-up (RR 2.68)
Despite elevated relative risks, 41–85% of patients with CDR 0.5–1 still pass on-road testing (100% of CDR 0 patients passed in the AAN-reviewed evidence). CDR alone cannot categorically revoke driving privileges — it identifies a higher-risk population, not an individual cut-off. Supplement with caregiver reports, crash/near-miss history, and formal on-road assessment. See Fitness to drive.

Aetiological subtype

  • Specify the presumed cause — Alzheimer's disease, frontotemporal lobar degeneration, Lewy body disease, vascular disease, traumatic brain injury, substance/medication use, HIV infection, prion disease, Parkinson's disease, Huntington's disease, another medical condition, multiple aetiologies, or unspecified.
  • Most subtypes carry a probable versus possible distinction, generally turning on whether there is genetic evidence or a sufficiently clear and characteristic clinical course. Say which you mean; "probable" and "possible" are not interchangeable in a letter that will follow the patient.
  • Multiple aetiologies is the honest answer more often than it is given. Mixed Alzheimer and vascular pathology is the commonest finding at autopsy in this age group; a formulation that names both is usually closer to the truth than one that forces a single label.

Behavioural specifier

Record with or without behavioural disturbance, and name the specific symptoms — psychotic symptoms, mood disturbance, agitation, apathy, or other. This is not a formality: it determines much of step 10, and it is the part of the diagnosis carers most need written down.

Writing the formulation

"Major neurocognitive disorder, moderate severity (CDR 2), due to probable Alzheimer's disease with a contributing vascular component, with behavioural disturbance (apathy and evening agitation). Supported by a three-year insidious amnestic decline with impaired delayed recall unhelped by cueing, MoCA 17/30, loss of independence in finances, medications and transport, MTA 3 with Fazekas 2 on MRI, and no delirium or alternative psychiatric explanation. Contributors identified and being addressed: untreated hearing loss, oxybutynin, and poorly controlled hypertension."

A formulation written this way answers, in one paragraph, every question the next clinician will have — and makes explicit which parts are firm and which are provisional.

10 ManagementDisclosure · non-pharmacological · medications · BPSD and PITCHED · follow-up · ACP

Nothing on the pharmacological list below matters as much as the first three items on the non-pharmacological one. Plan the management as a package with a named review date, not as a prescription.

Disclosing the diagnosis

  • Ask what they already suspect, and what they want to know, before you tell them. Most people want the diagnosis; a minority do not, and that preference deserves respect.
  • Use the word. "Dementia" or "Alzheimer's disease", not "memory problems" — ambiguity denies people the chance to plan while they still can.
  • Pair the diagnosis with what can be done: treatable contributors, supports, legal and financial planning, and the follow-up date. A diagnosis given without a plan is the version people remember badly.
  • Offer a written summary and a second appointment. Very little of the first conversation is retained.

Non-pharmacological management — first, and continuing

  • Treat every contributor found in step 6 — hearing and vision, sleep apnoea, depression, pain, deprescribing anticholinergics and sedatives.
  • Vascular risk management — blood pressure, diabetes, lipids, AF, smoking. Relevant to every subtype, not only vascular dementia.
  • Physical activity — the intervention with the best evidence across cognition, mood, falls and function. Prescribe it specifically.
  • Cognitive stimulation therapy for mild–moderate disease; social engagement and structured routine.
  • Occupational therapy for home safety, task simplification, and assistive technology.
  • Carer education, training and support — carer strain predicts institutionalisation more strongly than the patient's cognitive score. Dementia Australia (National Dementia Helpline 1800 100 500), carer support groups, respite, and Carer Gateway.
  • My Aged Care assessment for home supports; social work referral for financial and accommodation planning.
  • Nutrition, dental and continence review; falls prevention; immunisation.

Medications

Dementia Medications Guide, compiled 18 Aug 2026 by Rhymie Yusoff, Consultant Geriatrician, from a clinical evidence summary (OpenEvidence) — cognitive-enhancing medications and BPSD medications, side by side.

Informational guide only — not a clinical protocol, care pathway, or practice guideline. PBS listings, criteria, and subsidy rules change periodically — always confirm against the current PBS schedule before prescribing.
Part 1 — cognitive-enhancing medications

Cholinesterase inhibitors (donepezil, rivastigmine, galantamine) and the NMDA-receptor antagonist memantine. PBS subsidy for all four agents requires specialist involvement in diagnosis and periodic cognitive-score review to justify continued treatment.

DrugIndication / stagePBS criteria (Australia)Key adverse effectsMonitoringPreferred when
Donepezil
oral, once daily
Mild–severe Alzheimer's; off-label in DLB, PDD, vascular dementia Diagnosis confirmed/initiated with a specialist; MMSE ≥10 required; continuation beyond 6 months needs ≥2-point MMSE or ≥4-point ADAS-cog improvement (if baseline MMSE >24) Nausea, diarrhoea, anorexia/weight loss, insomnia, vivid dreams, muscle cramps, urinary frequency, falls; rare bradycardia/heart block Baseline ECG; TFTs; reassess ADLs/behaviour and MMSE/ADAS-cog at 6 months; watch for a prescribing cascade from incontinence Preferred first-line overall — widest severity range, best long-term observational data (slower decline, reduced mortality, delayed nursing-home placement)
Rivastigmine
oral or patch
Mild–moderate Alzheimer's (patch also approved for severe in some jurisdictions); PDD Same specialist-initiation and MMSE/ADAS-cog continuation rules as donepezil GI effects (patch ~3-fold less than oral); skin irritation at patch site (~1 in 10); muscle cramps, insomnia, urinary urgency Same as donepezil, including baseline ECG; rotate patch site Preferred in DLB/PDD — strongest evidence there; patch preferred with oral GI intolerance, swallowing difficulty, or adherence concerns
Galantamine
oral
Mild–moderate Alzheimer's; not indicated for mild AD alone Same PBS specialist-initiation; MMSE 10–14; continuation beyond 6 months requires ≥2-point MMSE improvement Similar cholinergic effects; dose reduction needed in severe renal impairment Same monitoring as other cholinesterase inhibitors Considered when other cholinesterase inhibitors are contraindicated/not tolerated; combination with memantine preferred over monotherapy in moderate–severe disease when tolerability allows
Memantine
NMDA antagonist
Moderate–severe Alzheimer's PBS-listed for moderate–severe AD; no indication for mild AD Dizziness, headache, constipation Renal function; MMSE at 6 months Preferred when agitation/aggression coexists with moderate–severe AD already warranting memantine — avoids an additional agent
Part 2 — BPSD medications

Behavioural and psychological symptoms of dementia — agitation, aggression, psychosis. Non-pharmacological strategies should generally be tried first; these agents carry meaningful risks in this population and most are used off-label with modest evidence of benefit.

AgentTarget symptomsAustralian use notesKey adverse effectsAvoid inPreferred when
RisperidoneAggression, agitation, psychosisOnly antipsychotic licensed in Australia for severe BPSD; review every 1–3 monthsExtrapyramidal symptoms, drowsiness, hypotension, hyperglycaemia, cerebrovascular events, fallsSuspected Lewy body dementia or Parkinson's diseaseSole agent with TGA/PBS approval and best supporting trial data when pharmacotherapy is genuinely needed
Haloperidol emergent use onlySevere emergent agitation (IM)Not recommended for routine BPSD in AustraliaProminent EPS, sedation, anticholinergic effects; highest mortality risk among antipsychotics in observational dataLewy body dementia/Parkinsonism; routine/chronic useReserved solely for acute emergent sedation when oral options aren't feasible
Quetiapine, olanzapine, aripiprazole off-labelAgitation, psychosisNot PBS-approved for BPSD specificallySedation, metabolic effects, falls, increased cerebrovascular events and mortality (class effect)Avoid as first-line; use only if risperidone unsuitableQuetiapine sometimes preferred in DLB/Parkinson's for lower EPS liability, despite weaker efficacy data
Citalopram (SSRI) off-labelAgitationNot PBS-listed for BPSD; used off-labelQTc prolongation (dose-dependent), GI effects, hyponatraemiaCaution with other QTc-prolonging drugsPreferred non-antipsychotic option for agitation when avoiding antipsychotic-related cerebrovascular/mortality risk is a priority
Mirtazapine not preferredAgitationOff-labelSedation, weight gainNot effective for agitation per recent evidence synthesisMay still suit cases where sedation and appetite stimulation are separately desirable (e.g. poor oral intake with insomnia)
Benzodiazepines (e.g. oxazepam)Severe anxiety, acute agitationPRN for escalating behaviours only; oxazepam preferred (short half-life, simple metabolism)Worsened cognition, falls, sedation, paradoxical disinhibitionChronic use in dementia; caution in all elderly patientsReserved for severe/acute presentations only, not routine/chronic use
Valproate / carbamazepine not preferredAgitation/irritability (historically trialled)Not PBS-indicated for BPSDSedation, hepatotoxicity, drug interactionsAvoid for BPSD given lack of efficacyLargely superseded — should not be first choice for any BPSD indication
Non-opioid analgesicsAgitation, especially if pain-drivenConsider trial where pain is a plausible contributorGenerally favourable safety profilePreferred first pharmacological trial when pain may be driving agitation — largest effect size (0.48) among BPSD interventions in recent synthesis
Part 3 — anti-amyloid monoclonal antibodies

Donanemab (Kisunla) and lecanemab (Leqembi) both received TGA registration during 2025 for early symptomatic Alzheimer's disease. Neither is PBS-subsidised, so in practice these remain privately funded and specialist-centre delivered — relevant to a small minority of patients, but increasingly asked about in clinic.

  • Indication — mild cognitive impairment or mild dementia due to Alzheimer's disease, with amyloid pathology confirmed by amyloid PET or CSF. Not for moderate or severe disease.
  • APOE ε4 genotype matters. The TGA registered lecanemab for non-carriers and heterozygotes only, having judged that safety was not satisfactorily established in heterozygotes and that neither safety nor efficacy was established in homozygotes. Genotyping is therefore part of the workup, not an optional extra.
  • ARIA — amyloid-related imaging abnormalities, oedema (ARIA-E) and haemorrhage (ARIA-H) — is the principal risk, requiring a baseline MRI with SWI and repeated surveillance MRI through the first six months. Pre-existing cerebral amyloid angiopathy, multiple microbleeds and anticoagulation substantially raise the risk.
  • Administration — intravenous infusion fortnightly or four-weekly at a specialist centre, which is a genuine access barrier outside metropolitan areas.
  • Cost — of the order of $40,000–50,000 per year for the drug, and roughly $80,000–100,000 once infusion and monitoring are included, while unsubsidised.
  • Benefit — a slowing of decline on group measures over 18 months, not an improvement and not a halt. The gap between what the trials showed and what families hope for is where most of the consultation time goes.
Regulatory status, PBS decisions and eligibility criteria in this area are moving quickly. Confirm the current TGA product information and PBS status before advising any individual patient.

BPSD — behavioural and psychological symptoms

Behaviour is communication. Before reaching for a drug, describe the behaviour precisely, look for what is driving it, and change that. Antipsychotics in this population carry an increased risk of stroke and death, and the number needed to harm is uncomfortably close to the number needed to treat.

PITCHED — the reversible drivers, checked every time

A Queensland-familiar checklist for the physical and environmental causes of changed behaviour in cognitive impairment. Work through it before, and alongside, anything pharmacological — it doubles as a delirium precipitant screen for step 5.

PPainAssess regularly with a validated tool (PAINAD/Abbey where self-report is unreliable). Agitation is often untreated pain; a trial of regular simple analgesia has one of the largest effect sizes of any BPSD intervention.
IInfectionUrinary, respiratory, skin. Maintain skin integrity and oral care; avoid unnecessary catheters.
TThirstOffer preferred drinks frequently, keep a fluid balance, limit caffeine late in the day.
CConstipationMonitor bowels actively; regular aperients, fibre and movement rather than reactive treatment.
HHungerMeal support, oral and dental care, weight tracking, dietitian referral where intake is falling.
EEnvironmentProtect the sleep–wake cycle, reduce noise and clutter, personalise the space, use clear orientation cues, and make sure glasses and hearing aids are actually on.
DDrugsReview every medication, avoid polypharmacy and anticholinergics, and treat psychotropics and sedatives as a last resort with a defined review date.

A structured approach to a behaviour

  • Describe — what exactly happens, when, where, with whom, how often, and who is distressed by it. Ask the staff or carer to keep a short behaviour chart for a week.
  • Investigate — work PITCHED; consider unmet need, boredom, fear, over- or under-stimulation, and the approach being used to deliver care.
  • Create — a plan the people actually present can carry out, written down, with one or two changes at a time.
  • Evaluate — at a defined date. If it did not work, change it rather than adding to it.
  • Escalate to pharmacological treatment only where there is risk of harm or severe distress that has not responded — then start low, review within days to weeks, define a stopping point, and attempt withdrawal after 3 months of stability. Avoid antipsychotics entirely in suspected DLB or PDD.

Prescribing detail for each agent, including the Australian regulatory position, is in Part 2 — BPSD medications above.

Follow-up

WhenWhat to review
2–4 weeks after diagnosisWhat they took in, what questions have surfaced since, carer reaction, medication tolerance if started, and whether the supports referred to have actually engaged
3 monthsResponse and adverse effects of any cognitive-enhancing agent; behavioural symptoms; carer strain; progress of legal and financial planning
6 monthsFormal cognitive re-test where a PBS-subsidised agent needs continuation criteria documented; functional review; driving; falls; weight; deprescribing opportunities
Then 6–12 monthlyFunction and stage rather than score alone; behavioural symptoms; carer health; care needs and accommodation; advance care planning review; goals of care as the disease advances
Any acute changeTreat as delirium until proven otherwise (step 5) — a sudden decline in dementia is a presentation, not a progression

Also review at each visit: current medication list and anticholinergic burden, continence, nutrition and weight, falls, pressure areas in later disease, immunisations, and whether the carer is coping.

Advance care planning

Start it early — while the person still has capacity to make the documents that matter. The window closes quietly, and families are left without authority at exactly the point they need it. Queensland instruments:

DocumentWhat it doesStatus
Enduring Power of Attorney (Forms 2 and 3)Appoints an attorney for personal/health matters and/or financial matters. Health decisions take effect only once capacity is lost; financial powers can start earlier if specifiedLegally binding. Requires capacity; witnessed by a JP, commissioner for declarations, lawyer or notary
Advance Health Directive (Form 4)Records directions about future health care and treatment, including specific refusals, and can appoint attorneysLegally binding. Requires capacity; a doctor or nurse practitioner must complete the health section, plus the usual witnessing
Statement of Choices (Form A, or Form B where the person no longer has capacity)Records values, goals and preferences to guide substitute decision-makers. Paediatric and culturally adapted versions exist for First Nations peoplesNot legally binding — guidance, and often the most useful document in practice
Acute Resuscitation Plan (ARP)Queensland Health clinical form recording resuscitation and treatment limitations for the current episode of careA clinical order completed by the treating doctor, informed by the documents above
  • Raise it at diagnosis, not at crisis. Frame it as ordinary planning that everyone should do, not as a conversation about dying.
  • Document capacity for each instrument at the time it is made — capacity is decision-specific, and a dementia diagnosis alone never removes it. See Capacity assessment.
  • Revisit at each stage transition, and make sure the documents are actually accessible — uploaded to My Health Record, copies with the substitute decision-maker, the GP and the residential facility.
  • As the disease advances, shift the conversation to goals of care, hospital transfer preferences, and palliative care involvement. Advanced dementia is a terminal illness and deserves to be planned for as one.

Safety and risk, revisited at every review

  • Driving — a legal reporting obligation as well as a clinical one in Queensland. See Fitness to drive.
  • Finances — scam and elder-abuse vulnerability rises early; check who has account access and whether anything has changed.
  • Home safety — cooking and appliances, smoking, wandering and getting lost, medication self-administration, firearms.
  • Carer safety — ask directly about aggression, and about the carer's own health.

Full source documents

DSM-5-TR–aligned formulation algorithm and its companion one-page flowchart, prepared by Rhymie Yusoff, Consultant Geriatrician — the highlights above are drawn from these. Opens the full PDF in a new tab.

Neurocognitive Disorder — Diagnostic Algorithm (32 pages)PDF
Quick-reference flowchart (1 page)PDF
Neuroanatomy · interactive

Clinical anatomy and neuroradiology of dementia

Where each dementia subtype does its damage — on a rotatable brain built from a real template scan, and on the MRI planes you read a scan in. Pick a pattern and both views shade the regions it characteristically involves.

Building the model…
Drag to rotate · scroll to zoom · click a structure Swipe to turn · pinch to zoom · tap to identify

Explore anatomy

Rotate freely and click any structure to see what it normally does.

The surface is a real cortical isosurface of the ICBM152 (MNI152) template brain, parcellated with the Harvard–Oxford atlas; the deep structures are that atlas's own hippocampus, amygdala, thalamus and basal ganglia. It is a normal template, not a patient — the shading marks where a subtype characteristically strikes, it is not atrophy rendered on this brain. Real-world overlap between subtypes is substantial; see the neurocognitive disorder pathway.

Where it shows on the scan

Real T1 sections through the ICBM152 template, in radiological convention — the patient's right is on your left. Selecting a pattern tints the regions it characteristically involves, using the Harvard–Oxford parcellation of the same brain, plus its white-matter tissue map for the periventricular and deep white matter.

RL
Coronal at the temporal poles (y = +10). The plane the frontotemporal syndromes declare themselves in — temporal pole and orbitofrontal cortex, neither of which a hippocampal-level coronal reaches.
RL
Coronal at the hippocampal body (y = −20). The MTA-scoring plane: choroid fissure, temporal horn and hippocampal height.
RL
Axial through the basal ganglia and thalamus (z = +6). Lobar atrophy, ventricular size and white-matter burden.
A
Sagittal near midline (x = −6). Precuneus and posterior cingulate for the Koedam scale, plus corpus callosum and midbrain.
Read every dementia scan in the same order. 1. Exclude the surgically treatable — subdural, tumour, normal pressure hydrocephalus. 2. Score the atrophy pattern with the visual rating scales. 3. Assess the vascular burden — white matter change, lacunes, strategic infarcts, microbleeds. 4. Only then look for the features specific to a subtype. Taking them in this order stops a striking atrophy pattern from distracting you off a treatable lesion.

Normal anatomy

No pattern selected — the sections are unshaded.

Reading the scan: visual rating scales

The four scales that turn "some atrophy" in a radiology report into something you can act on. Each is collapsed below with its grade definitions, the graded images, and the cut-off for calling it abnormal.

How the graded images were made, and what that means. Every grade strip below is the same ICBM152 template brain, with the change the grade is defined by applied to it — the hippocampus eroded and perihippocampal CSF widened for MTA, sulci widened for posterior and global atrophy, hyperintensities added on a synthesised FLAIR for Fazekas. They are simulations on real anatomy, not patient scans, so they show you what each grade changes with everything else held identical, which no real case series can. Use them to learn the feature; calibrate against real reported cases.
Medial temporal atrophy — Scheltens MTA (0–4)

Scored on coronal T1, angled perpendicular to the long axis of the hippocampus, at the level of the hippocampal body / anterior pons — and scored separately for each side. Three features move together: width of the choroid fissure, width of the temporal horn, and height of the hippocampus.

Real coronal T1 section through the medial temporal lobes
The plane itself, magnified — ICBM152 template, coronal at y = −22. These are the three structures the grade is read from.
Grade 00
Grade 0All three normal.
Grade 11
Grade 1Choroid fissure widened only.
Grade 22
Grade 2Temporal horn widened as well.
Grade 33
Grade 3Moderate loss of hippocampal height.
Grade 44
Grade 4Severe hippocampal volume loss.

Cut-offs — and why you will see two different sets

This is the part worth knowing properly, because the literature genuinely disagrees and both sets are in active clinical use.

AgeRadiology AssistantClassic EADC / ScheltensDecade-specific
(Ferreira 2015; Claus 2017)
<65≥1.5 — i.e. 2 on at least one side≥2 abnormal≥1.0–1.5
65–74≥1.5
75–84≥2.0 — i.e. 2 bilaterally≥3 abnormal≥2.0
≥85≥2.0–2.5 — no threshold performs acceptably
The half-point thresholds are not fudge — they encode how many sides are involved. The Radiology Assistant states it plainly: ≥1.5 means a 2 on at least one side, and ≥2.0 means a 2 bilaterally. That is the operational rule to standardise on, because it settles the question the raw numbers leave open. Two caveats stand regardless: score each side separately before you combine them, and above 85 the scale stops discriminating — every threshold there trades unacceptable sensitivity for unacceptable specificity.
Posterior atrophy — Koedam PA (0–3)

Assessed in all three planes — take the highest score. Sagittal: posterior cingulate sulcus, parieto-occipital sulcus, precuneus. Axial and coronal: posterior cingulate sulcus and parietal sulci. Best on 3D T1; axial FLAIR is a workable substitute.

Grade 00
Grade 0Sulci closed; no gyral atrophy.
Grade 11
Grade 1Mild widening, gyral bulk preserved.
Grade 22
Grade 2Substantial widening with gyral volume loss.
Grade 33
Grade 3End-stage "knife-blade" atrophy.

Cut-off: ≥2 is the rule-in threshold (sensitivity 58%, specificity 95% for AD). ≥1 is used instead when you want sensitivity — a deliberate trade, not a contradiction.

This is the scale that rescues young-onset AD. In pathologically confirmed AD, about 30% had posterior atrophy with a normal MTA; in early-onset AD, roughly a third had posterior atrophy only. In anyone under 65, or presenting with visuospatial, apraxic or dyscalculic problems, score PA before you take any comfort from a normal MTA.
Global cortical atrophy — GCA and the frontal subscale GCA-F (0–3)
GradeDefinition
0No cortical atrophy
1Mild — opening of sulci
2Moderate — volume loss of gyri
3Severe, end-stage — "knife-blade" atrophy
Grade 00
Grade 0No cortical atrophy.
Grade 11
Grade 1Mild — opening of sulci.
Grade 22
Grade 2Moderate — volume loss of gyri.
Grade 33
Grade 3Severe — "knife-blade" atrophy.

Best on axial FLAIR and/or 3D T1. GCA-F ≥1 is abnormal at all ages in the decade-specific work. The full Pasquier scale rates 13 regions separately for a 0–39 total; routine dementia reporting normally gives a single 0–3 impression instead.

GCA is the least reliable of the four. Trained memory-clinic raters reach κ ≈ 0.60 for GCA against ≥0.80 for MTA; in general non-specialist reporting, inter-rater agreement for total atrophy has been measured as low as κ 0.24 on MRI. Treat a single GCA number from a non-specialist report as soft.
Note: no published source gives an explicit slice/landmark rule for GCA-F — the validation work specifies only "the frontal sub-scale, on T1".
White matter hyperintensities — Fazekas (0–3)

Scored on axial FLAIR or T2. The widely-used combined form grades the overall burden: 0 none or very few punctate; 1 multiple punctate lesions; 2 beginning confluency — lesions bridging into each other; 3 large confluent lesions. The original scale rates periventricular and deep white matter separately (reported as a pair, e.g. "Fazekas 2/1"), which is still worth doing — deep change reflects small vessel disease more reliably, while periventricular change is heterogeneous in cause.

Grade 00
Grade 0PV: none. Deep: none.
Grade 11
Grade 1PV: caps or pencil-thin lining. Deep: punctate foci.
Grade 22
Grade 2PV: smooth halo. Deep: beginning confluence.
Grade 33
Grade 3PV: irregular, extending into deep white matter. Deep: large confluent areas.

Fazekas 1 is normal in an older person. 2 and 3 are pathological and carry meaningful risk of disability. The visual step-change that matters is confluence. Report the two separately: deep white matter change reflects small vessel disease more reliably, while periventricular change is heterogeneous in cause.

Imaging red flags — when the scan should change your diagnosis

These are the findings that stop a neurodegenerative work-up and send it somewhere else. Each is tied to the sequence it is actually best seen on, because several are easy to miss on the wrong one. This is the imaging arm of the Criteria D "exclude mimics" step.

These images show you where to look, not what the abnormality looks like. Each is the normal ICBM152 template with the relevant structure ringed and the measurement drawn on. I have deliberately not fabricated the pathology — a synthesised cortical ribbon or pulvinar sign would teach you to recognise my invention rather than the disease. Learn the location and the sequence here, then look at real reported cases for the appearance itself.
Normal pressure hydrocephalus

Evans index = A ÷ B ≥ 0.30, measured on the axial at the level of the maximal frontal horn width. Callosal angle is measured on the coronal, perpendicular to AC–PC at the posterior commissure: normal 100–120°, ≤90° abnormal, iNPH typically 50–80°. Look also for DESH — ventriculomegaly plus dilated Sylvian fissures plus effaced sulci at the high convexity.

Best sequence Axial and coronal T1 — the angle must be read on coronal

Sporadic CJD

Restricted diffusion in the cortical ribbon ± caudate and putamen. The classic criterion is caudate + putamen plus two cortical regions (sensitivity 69–76%, specificity 90–100%); a revised single-region criterion lifts sensitivity to 90–95%. Confirm true restriction on ADC — DWI shine-through is the standard trap.

Best sequence DWI + ADC — FLAIR is less sensitive and may be normal early

Variant CJD

The pulvinar sign — confluent hyperintensity of the pulvinar (posterior thalamus), symmetrical and brighter than the surrounding grey matter. When the dorsomedial nuclei light up with it the pair form the "hockey-stick" sign.

Best sequence Axial FLAIR; also visible on DWI

Autoimmune / limbic encephalitis

Bilateral T2/FLAIR signal change highly restricted to the medial temporal lobes — the wording of the Graus criteria. FDG-PET may be more sensitive than MRI, showing increased medial temporal uptake while the MRI still looks normal. On follow-up the resulting volume loss mimics AD.

Best sequence Coronal FLAIR/T2; FDG-PET if MRI is negative

Chronic subdural haematoma

A crescentic extra-axial collection that crosses suture lines and is minimally hyperintense to CSF on T1 and FLAIR. Internal membranes, loculations and mixed-age blood products mean rehaemorrhage. A subacute collection can be isodense on CT and disappear against adjacent cortex — which is exactly when it gets missed.

Best sequence FLAIR and T1 — not CT alone

Wernicke encephalopathy

Symmetric T2/FLAIR hyperintensity in the mammillary bodies, periaqueductal grey, tectal plate and dorsomedial thalami (the thalamic component sits higher, around z = +8). Mammillary body enhancement is common in acute disease. MRI sensitivity is only ~53% — a normal scan does not exclude it. Treat on suspicion, do not wait for the scan.

Best sequence FLAIR + DWI + post-contrast T1

Tumour

Dementia-mimicking tumours favour the middle cranial fossa, basal frontal region, thalamus, corpus callosum, cingulate gyrus and limbic system — sites where mass effect produces cognitive change before focal signs. CNS lymphoma: T2-hypointense, avidly enhancing, restricted diffusion.

Best sequence Post-contrast T1 + DWI

Patterns beyond the main subtypes

Four more that turn up in a cognitive clinic and have a signature worth knowing — two parkinsonian, two vascular. As above, these are the normal template with the relevant structure marked, not examples of the abnormality.

Progressive supranuclear palsy — the hummingbird

On midsagittal, the superior border of the midbrain is normally convex. In PSP it flattens and then becomes concave, and with the pons preserved behind it the outline reads as a hummingbird — midbrain the beak, pons the body. Look also for widening of the interpeduncular cistern. FDG-PET adds midbrain and medial frontal hypometabolism.

Best sequence Midsagittal T1

Multiple system atrophy

MSA-C: pronounced cerebellar and pontine atrophy, with the "hot cross bun" sign — cruciform T2 hyperintensity in the pons from degeneration of the transverse pontocerebellar fibres. MSA-P: dorsolateral putaminal hypointensity with a slit-like hyperintensity lateral to the putamen. Unlike PSP, the midbrain keeps its normal convex border — that is the single most useful discriminator between the two.

Best sequence Midsagittal T1 and axial T2 through the pons

Cerebral amyloid angiopathy

Lobar microbleeds at the corticosubcortical junction — peripheral, sparing the deep grey matter, which is the opposite of the deep distribution seen in hypertensive arteriolosclerosis. Add cortical superficial siderosis, lobar haemorrhage, and posterior-predominant white matter hyperintensities (Fazekas 2–3). This distribution changes anticoagulation decisions, so it is worth looking for deliberately.

Best sequence SWI or T2* — microbleeds are invisible on T1, T2 and FLAIR

CADASIL

Confluent white matter hyperintensities with anterior temporal pole involvement in close to 90% of cases — the single most useful feature separating it from sporadic small-vessel disease, which characteristically spares the temporal poles. External capsule involvement, lacunar infarcts and microbleeds complete the picture. Suspect it in early-onset vascular cognitive impairment, migraine with aura, or a family history of stroke.

Best sequence Coronal and axial FLAIR

Four things worth getting right

In DLB, the imaging findings everyone quotes are the supportive ones. Under the 2017 consensus criteria the only indicative biomarkers are reduced dopamine transporter uptake on SPECT/PET, reduced 123I-MIBG myocardial uptake, and polysomnographic REM sleep without atonia — these can lift a diagnosis to probable DLB. Medial temporal preservation, occipital hypometabolism and the cingulate island sign are supportive only: consistent with DLB, but they do not change the level of diagnostic certainty.
A normal MTA does not exclude Alzheimer's disease. Pooled performance of MTA for AD versus controls is roughly 74% sensitivity, 88% specificity — about one in four people with AD is missed by MTA alone, even under favourable conditions. Score posterior atrophy too.
A high MTA is not specific for Alzheimer pathology, and least so in the oldest patients. Medial temporal atrophy rises with age independently of AD; white matter hyperintensity burden is the strongest single non-AD predictor of it. And LATE (limbic-predominant age-related TDP-43 encephalopathy) is present in over 20% of people past 80, produces an amnestic syndrome that mimics AD clinically, and causes hippocampal atrophy that is greater than in pure AD.
DLB and Parkinson's disease dementia cannot be told apart on imaging. They share the same α-synuclein signature — preserved medial temporal lobes, posterior/occipital hypometabolism, abnormal DAT and MIBG. The distinction is temporal, not radiological: the one-year rule, dementia before or within a year of motor onset for DLB, dementia in established Parkinson's for PDD.

Sources for this section

  • Scheltens P, et al. J Neurol Neurosurg Psychiatry 1992
    Original medial temporal atrophy visual rating scale
  • Ferreira D, et al. J Intern Med 2015;278:277–90
    Practical decade-specific cut-offs for MTA, GCA-F and posterior atrophy
  • Claus JJ, et al. Eur Radiol 2017;27:1115–21
    Independent memory-clinic validation of decade-specific MTA cut-offs; the ≥85 problem
  • Koedam ELGE, et al. Eur Radiol 2011;21:2618–25
    Posterior atrophy visual rating scale and its cut-off
  • Lehmann M, et al. NeuroImage Clin 2013
    Posterior atrophy without medial temporal atrophy in pathologically confirmed AD
  • Fazekas F, et al. AJR 1987
    White matter hyperintensity grading
  • McKeith IG, et al. Neurology 2017;89:88–100
    Fourth consensus report of the DLB Consortium — indicative vs supportive biomarkers
  • Rascovsky K, et al. Brain 2011;134:2456–77
    International consensus criteria for behavioural-variant FTD, including imaging criteria
  • Gorno-Tempini ML, et al. Neurology 2011;76:1006–14
    Classification of primary progressive aphasia and its variants, including imaging criteria
  • Nelson PT, et al. Brain 2019;142:1503–27
    LATE consensus working group report
  • Duering M, Wardlaw JM, et al. Lancet Neurol 2023;22:602–18
    STRIVE-2 neuroimaging standards for small vessel disease
  • Sachdev P, Bentvelzen A, et al. JAMA Neurol 2025
    VasCog-2/WSO revised diagnostic criteria for vascular cognitive impairment
  • Minoshima S, et al.; Lim SM, et al. J Nucl Med 2009;50:1638–45
    Occipital hypometabolism and the cingulate island sign in DLB
  • The Radiology Assistant — Dementia: role of MRI (Barkhof, Sanchez, Hoogenboom)
    Grade definitions, MTA age cut-offs and the four-step reading order used throughout this section — radiologyassistant.nl/neuroradiology/dementia/update
The slices and grade illustrations on this page are hand-drawn schematics, not images derived from any patient scan or published atlas. They are built to teach where to look and what changes between grades; they are not a reference standard, and no diagnosis should rest on a comparison against them. Where the literature genuinely disagrees — most importantly the MTA cut-offs — both positions are shown rather than one being picked for you.
Decision-making capacity

Capacity assessment

Capacity is decision-specific and time-specific — a patient can lack capacity for one decision while retaining it for another. Model framework: Queensland Powers of Attorney Act 1998 and Guardianship and Administration Act 2000. Check local legislation outside Queensland.

The four-part test

This test explains what's being judged — it is not something a junior doctor administers themselves.

  • 01

    Understand

    Understand the information relevant to the decision.

  • 02

    Retain

    Retain that information long enough to use it.

  • 03

    Weigh up

    Weigh it up as part of the decision-making process.

  • 04

    Communicate

    Communicate the decision in some way.

WhoRole
ConsultantPerforms and formulates formal capacity assessments and cognitive diagnoses; leads QCAT referrals
Advanced traineeMay perform these assessments, always under direct consultant supervision
Junior doctorNot expected to assess capacity or formulate a cognitive diagnosis — gather collateral history, note observations, and escalate concerns

Decision domains

Capacity is decision-specific — a finding in one domain does not transfer to another. Each domain below is assessed against its own legal test.

Personal / lifestyle

Day-to-day personal care, social contact, and lifestyle choices short of a permanent move — usually the domain retained longest.

Health care

Consenting to or refusing medical treatment and investigations — assessed per decision, and per the standard four-part test for that specific treatment.

Financial

Managing income, bills, assets, and financial decisions — often the first domain to decline, and frequently under-recognised until a crisis (unpaid bills, scams).

Accommodation

Deciding on a permanent change in living arrangements, such as a move into residential aged care — a materially higher-stakes decision than day-to-day personal care.

Enduring Power of Attorney (EPOA)

Capacity to appoint an attorney for personal and/or financial matters — a lower bar than the underlying decision itself, since the person is choosing someone they trust, not making the decision.

Testamentary capacity

Capacity to make or alter a will — classically the Banks v Goodfellow test: understanding the nature of making a will, the extent of one's property, and the claims of those who might expect to benefit.

Advance health directive

Capacity to complete or amend a document giving direction about future health care, ahead of losing capacity to decide at the time.

Other domains

Marriage, voting, sexual consent, entering contracts, instructing a lawyer, and similar — each has its own specific legal test; there is no single "global" capacity.

Capacity assessment register

An interactive companion for assessing decision-specific capacity across ten domains — including capacity to appoint an Enduring Power of Attorney.

Capacity Assessment Register self-authored reference · QLD framework
General reference tool only — not legal or medical advice, and not a determination of capacity. Nothing entered here is saved, transmitted, or stored — state lives only in this browser tab for the session. Do not type patient names, MRNs, or other identifiers.
Austroads Assessing Fitness to Drive, 2022 edition

Fitness to drive

A searchable quick reference distilled from the Austroads medical standards for licensing, covering both private and commercial vehicle drivers. CDR-based dementia driving risk is covered alongside the broader neurocognitive disorder pathway.

Fitness to Drive — Quick Reference adapted from Austroads (2022)
A quick-reference distillation, not a replacement for the full Austroads standard or for AP-G56-22 (kept alongside this hub) when a reporting obligation or borderline case needs the primary source.
Bedside consultation tool

Geriatric clinic companion

History · exam · investigations · formulation · plan — an interactive companion to the Comprehensive Geriatric Clinic Guide and the NCD Diagnostic Algorithm, built to run live during a consultation.

Geriatric Clinic Companion self-authored reference
Nothing here is saved, transmitted, or stored anywhere — state lives only in this browser tab for this session and clears on reload. Do not type patient names, MRNs, or other identifiers.
Teaching & provenance

Education

Ward-round teaching topics, a flash-card deck built from the rest of the hub, and the evidence base and disclaimer that apply to every page here.

Grand round roulette

Spin to pick what you'll teach next — weighted toward geriatric and general medicine, with procedures, communication, wellbeing, and training mixed in. Made for ward-round and tutorial teaching.

Grand Round Roulette self-authored teaching tool

Flash cards

Ninety-four questions drawn from the pages of this hub — every answer is what the hub itself says, so the deck and the reference cannot drift apart. Pick your decks, choose whether you want to test yourself by flipping or by choosing, and any card will take you back to the passage it came from.

Flash card deck 94 questions drawn from this hub
Decks
Level
Mode
Round

Comprehensive geriatric assessment

Six connected domains — physical, psychological, functional, social, environmental, and future wishes — laid over a single home scene. Hover or tap a person or an object to open that domain's prompts. The point it keeps making is that CGA is not a stack of separate assessments but one prioritised problem list and care plan the team builds together.

Interactive Comprehensive Geriatric Assessment self-authored teaching tool

PITCHED — searching for delirium triggers

Pain, Infection, Thirst, Constipation, Hunger, Environment and Drugs, laid over an acutely confused patient on a ward. Hover or tap each cue for what to ask and what to do about it. It searches for contributors rather than making the diagnosis — confirm acute change and inattention with 4AT or CAM first.

PITCHED Delirium Assessment self-authored teaching tool

Geriatric medicine body map

Ten body regions, each opening the geriatric syndromes that belong to it — brain and behaviour through to feet. The internal anatomy stays hidden until the region above it is explored.

Geriatric Medicine Body Map self-authored teaching tool

SPDP

Reserved for the supervised practice and professional development plan material.

🚧 Under construction

Nothing here yet — this section is a placeholder while the content is written.

Evidence base

  • DSM-5-TR (2022)
    Neurocognitive Disorders diagnostic criteria — the diagnostic framework underlying the NCD pathway
  • Clinical Practice Guidelines and Principles of Care for People with Dementia in Australia
    Cognitive Decline Partnership Centre / NHMRC-endorsed
  • NICE NG97; AAN practice guidelines
    Guideline synthesis for cognitive assessment and driving-risk correlates
  • Consensus aetiological diagnostic criteria
    NIA-AA/IWG-2 (Alzheimer disease) · McKeith 2017 (Lewy body disease) · Rascovsky 2011 (behavioural-variant FTD) · Gorno-Tempini 2011/2016 (primary progressive aphasia) · Graus 2016 (autoimmune encephalitis) · WHO/CDC criteria (Creutzfeldt–Jakob disease)
  • Johnson JCS, McWhirter L, Hardy CJD, et al. Suspecting dementia: canaries, chameleons and zebras.
    Pract Neurol 2021;21:300–312 — the symptom-led bedside framework for atypical presentations, incorporated with permission into the NCD algorithm's bedside sections
  • Austroads. Assessing Fitness to Drive for Commercial and Private Vehicle Drivers, 2022 edition.
    Medical standards for licensing and clinical management guidelines — source for the Fitness to Drive quick reference
  • Queensland Powers of Attorney Act 1998 & Guardianship and Administration Act 2000
    Model legislative framework for the capacity assessment content — check local legislation outside Queensland
  • OpenEvidence clinical evidence summaries
    Basis for the geriatric screening bloods and dementia medications comparison tables

What's original vs. adapted

Every interactive tool embedded in this hub, the NCD diagnostic algorithm, its quick-reference flowchart, and the medications guide were compiled and authored by Rhymie Yusoff, Consultant Geriatrician, GEM/Rehabilitation Unit, Logan Hospital, Metro South Health — synthesising the published sources above for personal practice and junior doctor teaching. The fitness-to-drive quick reference adapts the Austroads standard; the full 264-page Austroads document and the cited journal literature remain the primary sources where a reporting obligation or borderline case requires them.

This hub is a personal clinical reference and teaching aid, not a validated protocol, institutional policy, or formally endorsed practice guideline. It is a structured synthesis for individual practice and must be read alongside — never in place of — current published guidelines, local institutional protocols, and individual clinical judgement for each patient. PBS/MBS criteria and legislative frameworks change periodically — always confirm against current sources before clinical or legal use.