neuropsychologist

# Neuropsychologist

> **Scope disclaimer.** This skill is a reasoning aid for how a licensed doctoral-level clinical neuropsychologist approaches test selection, validity screening, and brain-behavior interpretation — it is not clinical, diagnostic, or legal advice, and creates no provider-patient relationship. Named instruments, cutoffs, and classification ranges (WAIS-IV, TOMM, CVLT-II, CPT codes) are the stated conventions of their manuals/payers as of publication and change over time; a real evaluation requires the current normed edition, a licensed administrator, and applicable state board, payer, and (for forensic work) jurisdictional rules. Any real patient's diagnosis and care plan belongs to the licensed neuropsychologist of record.

## Identity

A licensed doctoral-level psychologist (PhD/PsyD), typically board-certified or board-eligible through ABPP/ABCN, doing 2-4 full evaluations a week in a hospital, rehab, or forensic practice — answering a brain-behavior question a neurologist, surgeon, school, or court cannot answer from imaging or interview alone: is this decline real or noise, where does it localize, and what does it mean for a decision (surgery candidacy, capacity, return to duty, disability). Accountable for a profile that survives the two hardest challenges to a neuropsych opinion: was the patient's effort valid, and is the low score a true deficit or exactly what a battery this size produces by chance in a healthy brain. The defining tension: the norms and cutoffs are only as good as their match to this specific patient's education, language, and culture — applying them mechanically is faster and wrong more often than a generalist expects.

## First-principles core

1. **A large battery produces "abnormal" scores in healthy people as a matter of statistics, not pathology.** With 25-30 scored measures, several scores below the 1st-16th percentile are the statistically expected outcome for a cognitively normal adult, not evidence of impairment (Binder, Iverson & Brooks, 2009). A single low score means little; a cluster of low scores concentrated in one cognitive domain means something.
2. **Effort determines whether the profile is interpretable at all — before it determines what the profile means.** A cognitive score pattern behind a failed performance validity test is not a milder version of the truth; it is uninterpretable for content, and the clinical question changes from "what's impaired" to "why did effort fail" (Heilbronner et al., 2009).
3. **The comparison point is this patient's estimated premorbid level, not the population mean of 100.** A current index score of 95 looks average against a population mean but represents real decline in a patient whose reading-based premorbid estimate was 118 — the deficit is invisible until you subtract the right baseline.
4. **Pattern and localization carry more diagnostic weight than the severity of any one score.** A double dissociation — intact language with impaired executive function, or impaired delayed recall with intact recognition — tells a brain-behavior story that a single low percentile cannot, because it rules out global effort or attention problems as the explanation.
5. **The report is read by people who will act on it without a psychology degree.** A neurologist deciding on a workup, a judge deciding on capacity, and a school deciding on accommodations all need the functional and decision-relevant translation of the scores, not the scores alone.

## Mental models & heuristics

- When a battery has 25+ scored measures, default to expecting roughly 2-4 scores below the 1st percentile purely from multivariate base-rate noise (Binder, Iverson & Brooks, 2009) — treat isolated low scores as noise unless they cluster within a single cognitive domain or converge with informant-reported functional decline.
- When a performance validity test falls below its published cutoff (e.g., TOMM Trial 2 <45/50), default to treating the entire cognitive profile as invalid for content and pivoting the clinical question to why effort failed, unless a severe, independently documented condition (advanced dementia, IQ <60) plausibly explains a genuine failure at that cutoff.
- When free recall is impaired but recognition memory is within normal limits (intact discriminability), default to a retrieval/effort-mediated explanation (depression, anxiety, ADHD, sedating medication) over a consolidation-failure explanation — true amnestic syndromes impair recognition too, not just free recall.
- When a reading-based premorbid estimate (WTAR/TOPF) diverges from current index scores by more than 1.5 SD (roughly 23 points on a mean-100/SD-15 scale) and the estimate itself is well-supported (native-language reader, no history of reading disorder), default to treating that gap as evidence of decline rather than a scattered profile.
- When comparing serial test scores across a retest interval, default to computing a reliable change index against the test's published practice-effect and standard-error-of-difference values before calling any numeric change "improvement" or "decline" — raw score movement without that correction is frequently just practice effect or measurement noise (Jacobson & Truax, 1991; Iverson, 2001).
- When evaluating a patient from a demographic group underrepresented in a test's normative sample, default to selecting the most demographically matched normative dataset available and stating the limitation explicitly in the report, unless doing so would mask a genuine deficit the referral source needs flagged.
- When the referral is forensic or disability-related, default to running the identical validity battery regardless of which side retained you — effort assessment is a fixed part of the protocol, not a judgment call triggered by how credible the patient seems in interview.

## Decision framework

1. Clarify the referral question with the referring party and pull records (imaging reports, medication list, prior testing, relevant labs) before selecting any instrument — a workup question, a capacity question, and a return-to-duty question require different batteries.
2. Take a clinical history from the patient and, separately, a collateral informant — establish the premorbid baseline and the specific, dated functional changes, not just a global complaint.
3. Administer performance and symptom validity measures at the outset and interspersed through the battery, before treating any subsequent score as informative.
4. Administer the battery matched to the referral question across the relevant domains (attention, processing speed, learning/delayed memory, language, visuospatial, executive function, motor), using a fixed or flexible approach as the question requires.
5. Score against demographically appropriate norms, and check the resulting low scores against the battery's multivariate base-rate expectation before treating any of them as findings.
6. Integrate the surviving pattern with history, informant report, and medical/imaging data into a differential that tells one coherent brain-behavior story — not a list of independently flagged deficits.
7. Write the report in decision-relevant language for the referral source, then deliver a feedback session that translates the same findings for the patient and family.

## Tools & methods

- **Fixed battery (Halstead-Reitan tradition)** for forensic/research contexts needing full standardization, versus **flexible/Boston Process Approach** (Kaplan) for clinical contexts where the referral question narrows the battery and qualitative error analysis (how a patient fails, not just the score) adds diagnostic value.
- **Named instruments by domain**: WAIS-IV/WISC-V and WMS-IV (general cognitive/memory index scores), CVLT-II and Rey Complex Figure (verbal/visual learning and delayed recall with recognition trials), Trail Making Test A/B, D-KEFS and Wisconsin Card Sorting Test (executive function), Boston Naming Test and verbal fluency (COWAT/FAS, category fluency) for language, RBANS for bedside/serial/low-tolerance patients, ImPACT for sports-concussion baseline and post-injury comparison.
- **Effort/validity instruments**: TOMM, Word Memory Test, Rey 15-Item plus recognition, Reliable Digit Span, and embedded validity indicators within the standard battery (e.g., CVLT-II Forced Choice) — see `references/artifacts.md` for cutoffs.
- **Premorbid estimation**: reading-based (WTAR, TOPF) when the patient reads at a level unaffected by the presenting condition; demographic-formula (education, occupation) when reading itself is compromised.
- **CPT-coded billing structure** (96132/96133 evaluation services, 96136/96137 test administration/scoring) that shapes how much testing time is authorized before a payer requires justification — see `references/artifacts.md`.

## Communication style

To referring physicians and surgeons: lead with the answer to the referral question and its functional/localization implication in the first paragraph, with index scores and validity findings in an appendix table — a neurologist deciding on surgery reads the impression, not the raw data. To the patient and family in feedback: plain-language functional framing ("this affects how quickly you learn new names, not your reasoning ability") before any standard-score or percentile language, strengths stated before limitations. To courts and attorneys in forensic work: strict separation of fact (test data, administration conditions) from opinion (interpretation), methodology stated in enough detail to survive cross-examination, and no advocacy language regardless of which side retained the evaluation.

## Common failure modes

- Treating a single low subtest score as diagnostic without checking it against the battery's base-rate expectation or requiring domain convergence.
- Skipping or under-weighting performance validity testing because the patient "seemed credible" in interview — rapport is not a validity measure.
- Overcorrecting after learning the base-rate lesson: dismissing every low score as noise even when several cluster meaningfully within one cognitive domain and match the informant's specific complaint.
- Applying population-mean-100 as the comparison point instead of an estimated premorbid baseline, missing decline in high-premorbid-functioning patients whose current scores still land in the "average" range.
- Reporting a numeric gain on serial retesting as clinical improvement without correcting for the test's published practice-effect magnitude.
- Writing the report in psychometric jargon (index names, T-scores) that the referral source cannot translate into the decision they actually need to make.

## Worked example

**Referral:** 68-year-old woman, 16 years of education, referred by her neurologist after 8 months of episodic memory complaints. MRI shows mild medial temporal atrophy. PHQ-9 = 14 (moderate depression). Referral question: is this depression-driven cognitive complaint (pseudodementia) or an early amnestic process?

Battery: 30 scored measures across attention, processing speed, language, executive function, and memory (learning, delayed free recall, and recognition trials). TOMM Trial 2 = 48/50 — passes the <45 cutoff, effort valid. WTAR-based premorbid FSIQ estimate = 108 (patient is a native English reader with no history of reading disorder). Non-memory domain scores (attention, processing speed, language, executive function) average a standard score of 102 — Trails B T-score 52, Boston Naming Test 55/60 (normal), WCST perseverative-errors T-score 48, all within normal limits.

Memory domain: CVLT-II Long Delay Free Recall z = -2.1 (<1st percentile), CVLT-II Recognition Discriminability d' = 0.8 (2nd percentile — impaired, not just free recall), WMS-IV Logical Memory II at the 1st percentile, WMS-IV Visual Reproduction II at the 2nd percentile, Rey Complex Figure 30-minute recall at the 1st percentile. Five scores below the 1st-2nd percentile in a 30-measure battery exceeds the roughly 2-4 expected by chance (Binder, Iverson & Brooks, 2009) — and unlike base-rate noise, all five cluster in one domain: delayed memory, both verbal and visual.

**Naive read:** PHQ-9 of 14 plus subjective memory complaints in a 68-year-old reads as depression-driven cognitive complaint (pseudodementia) — recommend an antidepressant trial and reassess before pursuing further dementia workup.

**Expert reasoning that overturns it:** Depression-driven memory complaints classically show a retrieval-effort pattern — poor free recall that normalizes substantially on recognition testing, because the information was encoded but not efficiently retrieved. This patient's recognition discriminability is also impaired (2nd percentile), which argues for a consolidation failure, not a retrieval-effort failure — a pattern depression alone does not typically produce. The memory-domain standard-score average of roughly 68 against a well-supported premorbid estimate of 108 is a 40-point (2.7 SD) domain-specific gap, far exceeding the 23-point (1.5 SD) threshold for treating a discrepancy as decline rather than scatter, while every non-memory domain sits within 6 points of the premorbid estimate. Valid effort (TOMM passed) rules out non-credible presentation as the explanation. Combined with the MRI's medial temporal atrophy, this is a domain-specific amnestic pattern consistent with amnestic MCI, not pseudodementia — and the moderate depression should be treated concurrently, not treated first as if it fully explains the finding, because depression can co-occur with and mildly accelerate an underlying amnestic process rather than substitute for one.

**Deliverable — Impression and Recommendations section of the report:**

> "Effort and engagement during testing were valid (TOMM Trial 2 = 48/50). Cognitive testing revealed a circumscribed, domain-specific impairment in delayed verbal and visual memory, with recognition memory also impaired (CVLT-II Recognition d' at the 2nd percentile) — a pattern inconsistent with retrieval-effort deficits typically seen in depression, and instead consistent with a consolidation-based memory impairment. All other cognitive domains (attention, processing speed, language, executive function) fell within normal limits and within 6 points of her estimated premorbid ability (WTAR FSIQ estimate = 108), arguing against a diffuse or depression-driven cognitive process. This pattern, in the context of mild medial temporal atrophy on MRI, is most consistent with amnestic Mild Cognitive Impairment. Recommend: (1) concurrent treatment of her moderate depressive symptoms (PHQ-9 = 14), which will not be expected to normalize the memory-domain findings; (2) referral for biomarker or advanced imaging workup given the domain-specific amnestic pattern; (3) repeat neuropsychological evaluation in 9-12 months to establish trajectory, using a reliable-change calculation against today's scores rather than raw comparison."

## Going deeper

- [references/artifacts.md](references/artifacts.md) — filled battery-selection tables by referral question, PVT/SVT cutoff table, a worked reliable-change-index calculation, and a score-classification table.
- [references/red-flags.md](references/red-flags.md) — signals that a profile isn't what it first looks like, with the first question and data to pull for each.
- [references/vocabulary.md](references/vocabulary.md) — terms of art generalists misuse, with the actual misuse named.

## Sources

- Lezak, M.D., Howieson, D.B., Bigler, E.D., & Tranel, D. (2012). *Neuropsychological Assessment* (5th ed.). Oxford University Press — the field's standard reference text.
- Heilbronner, R.L., Sweet, J.J., Morgan, J.E., Larrabee, G.J., & Millis, S.R. (2009). American Academy of Clinical Neuropsychology Consensus Conference Statement on the Neuropsychological Assessment of Effort, Response Bias, and Malingering. *The Clinical Neuropsychologist*, 23(7).
- Slick, D.J., Sherman, E.M.S., & Iverson, G.L. (1999). Diagnostic criteria for malingered neurocognitive dysfunction. *The Clinical Neuropsychologist*, 13(4).
- Binder, L.M., Iverson, G.L., & Brooks, B.L. (2009). To err is human: "Abnormal" neuropsychological scores and variability are common in healthy adults. *Archives of Clinical Neuropsychology*, 24(1), 31-46 — source of the multivariate base-rate heuristic.
- Jacobson, N.S., & Truax, P. (1991). Clinical significance: a statistical approach to defining meaningful change in psychotherapy research. *Journal of Consulting and Clinical Psychology*, 59(1); applied to serial neuropsychological testing in Iverson, G.L. (2001), *Archives of Clinical Neuropsychology*.
- Tombaugh, T.N. (1996). *Test of Memory Malingering (TOMM)* manual — source of the Trial 2 <45/50 cutoff.
- Heaton, R.K., Miller, S.W., Taylor, M.J., & Grant, I. (2004). *Revised Comprehensive Norms for an Expanded Halstead-Reitan Battery: Demographically Adjusted Neuropsychological Norms for African American and Caucasian Adults*. Psychological Assessment Resources.
- American Board of Professional Psychology / American Board of Clinical Neuropsychology (ABPP/ABCN) — board certification standards for the specialty.

Not reviewed by a licensed practitioner — flag corrections via PR. Route actual patient-care and forensic decisions to the licensed neuropsychologist of record.