How to Read a CPT Report: A Clinician Walkthrough

A continuous performance test (CPT) report can look like a wall of numbers — T-scores, percentiles, error counts, reaction-time statistics — and it’s easy to either over-read a single red flag or skim past what the report is actually telling you. Reading one well means working through a consistent sequence: check validity first, then read each metric domain (attention, impulsivity, consistency, vigilance over time), then weigh the whole pattern against the interview and rating scales, not any one number in isolation. This walkthrough covers the metrics that show up across most commercial CPTs, using Conners CPT 3 and T.O.V.A. terminology as reference points since they’re the most documented.

Photo by Sasun Bughdaryan on Unsplash

Step 1: Check that the result is valid before you interpret it

Before reading any score, confirm the test was completed under conditions that make the numbers meaningful. Most CPT platforms flag or report:

• Completion and consistency. Did the patient finish the full test? An interrupted or restarted session produces unreliable statistics.
• Response style indices. Some CPTs (including Conners CPT 3) include embedded validity indicators meant to flag unusually careless, random, or overly cautious responding, separate from the clinical scores [1].
• Environment. For remote tests, note whether the patient reported distractions, used an unfamiliar device, or tested at an unusual time relative to medication dosing. Our guide to remote CPT testing validity covers what the research says about environment effects and how to standardize them.

If validity flags are present, treat the scores that follow with more caution, or consider re-testing before making a clinical decision based on the result.

Step 2: Understand what a T-score or standard score actually tells you

Most CPT reports convert raw performance into a standardized score compared to an age- and sex-matched (or age-and-gender) normative sample, rather than reporting raw numbers. Conners CPT 3, for example, reports T-scores with a mean of 50 and a standard deviation of 10; scores above roughly 60 are considered atypical, with higher values reflecting a greater departure from the normative sample, and interpretable elevations generally start in the low 60s and become more pronounced above 70 [1]. T.O.V.A. instead reports several of its variables as standard scores on a different scale, where clinicians generally treat scores below about 80 as abnormal, 80–85 as borderline, and above 85 as within normal limits, alongside a separate ADHD-classification score [2]. The scale differs by product, so the first thing to check on any report is which normative metric it uses and what its cutoffs mean — don’t assume a “70” or a “90” means the same thing across two different CPTs. AxonCPT reports results against cohort-matched norms drawn from a reference dataset of more than 20,000 tests across 12 age × sex cohorts, so a given patient is compared to same-age, same-sex peers rather than a single pooled norm [4].

Step 3: Read omission and commission errors as two separate signals

Omission errors (missed targets) and commission errors (responses to non-targets) are the two most-cited metrics on almost any CPT report, and they map onto different symptom domains: omissions toward inattention, commissions toward impulsivity or weak inhibitory control. A patient with a high omission rate and a low commission rate is telling a different story than the reverse pattern, even with a similar total error count. Reading the two together with reaction time adds detail: a high commission rate paired with unusually fast reaction times is a classic impulsive-response pattern, while a high omission rate paired with reaction times that slow and grow more variable across the session points toward a sustained-attention problem. For a full breakdown of what each error type reflects clinically, see omission vs. commission errors in ADHD testing.

Step 4: Read reaction time and its variability, not just accuracy

Two other metrics round out most reports:

• Hit reaction time (mean RT). How fast the patient responds to targets, on average. Very fast RT paired with high commissions suggests impulsive responding; very slow RT paired with high omissions suggests disengagement.
• Reaction time variability (RT standard deviation). How consistent the patient’s response speed is across the test. High variability — a patient whose speed swings between very fast and very slow — is one of the more consistently reported findings in ADHD samples and is sometimes considered as informative as the error rates themselves.
• Detectability (d′). A signal-detection measure of how well the patient distinguishes targets from non-targets overall, independent of response bias; poor detectability plus a high error rate paints a more complete picture than either alone [1].

None of these numbers should be read as a pass/fail gate. They’re most useful as a pattern: does the whole profile — errors, speed, and consistency — point in a consistent direction, or is it a scatter of unrelated single flags?

Step 5: Look at the pattern across the session, not just the summary score

A single summary score can hide a patient who started strong and fell apart in the last third of the test, or one who was slow to warm up but stabilized. Where a report breaks results into quarters or blocks across the session — AxonCPT reports a Q1–Q4 vigilance breakdown for this reason [4] — that time-course view is often more clinically informative than the overall average, especially for questions about sustained attention specifically (as opposed to attention at any single moment). A patient whose omission rate climbs steadily from Q1 to Q4 is showing a vigilance decrement; a patient who is uniformly elevated throughout is showing a different pattern.

Try it with your next patient. AxonCPT gives you 2 free tests every month — no card required. Send a link, your patient finishes in ~16 minutes from home, and you get a norm-referenced report instantly. Create your free account →

Step 6: Weigh the whole report against the rest of the evaluation

This is the step that’s easiest to skip under time pressure, and the one the evidence says matters most. Pooled across commercially available CPTs, a 2024 meta-analysis in the Journal of the American Academy of Child & Adolescent Psychiatry found sensitivity of 0.75 and specificity of 0.71 for the overall score, concluding that no CPT metric reviewed met a 0.80/0.80 standard for acceptable standalone accuracy [3]. That means roughly a quarter of true ADHD cases will not show an abnormal result, and roughly three in ten people without ADHD will. A normal CPT result should be read as reassuring, not disqualifying, and an abnormal result as a prompt for closer clinical review, not a diagnosis. For the complete evidence picture behind that recommendation, see CPT accuracy for ADHD: what the meta-analyses show.

A few practical rules follow from that evidence:

• Weight a convergent pattern more heavily than any single elevated score. An abnormal commission-error T-score is weaker evidence alone than the same finding alongside a consistent history, informant report, and elevated rating-scale scores.
• Consider confounds. Anxiety, depression, poor sleep, and low effort can all shift CPT scores without reflecting a stable attention or impulse-control problem; if the clinical picture doesn’t fit the test result, investigate rather than defer to the number.
• Use repeat testing to track change, not to re-diagnose. Comparing the same patient’s scores across visits — before and after a medication change, for example — is a different and often more clinically useful application than single-administration classification, since the patient serves as their own baseline.

How this looks in practice

A psychologist or psychiatrist reviewing a CPT report as part of a full evaluation typically works through it in this order:

1. Check validity flags and completion status first, before reading any clinical score.
2. Note which normative scale the report uses (T-score, standard score, percentile) and what its cutoffs mean for that specific product.
3. Read omissions, commissions, reaction time, and variability together, looking for a convergent pattern rather than isolated elevations.
4. Check the time-course breakdown, if the report includes one, for whether attention declines, stays flat, or is uniformly elevated across the session.
5. Weigh the whole picture against the interview, history, and rating scales, and document how you integrated them.
6. For a follow-up patient, compare directly to the prior report rather than re-evaluating from zero. With objective ADHD testing with AxonCPT, each report is built for that kind of side-by-side comparison across visits, and you can see what the report looks like before your first test.

Frequently asked questions

What’s a “normal” CPT score? It depends on the product’s scale. On Conners CPT 3, T-scores near 50 are typical, with scores above roughly 60 flagged as increasingly atypical [1]. On T.O.V.A., standard scores above about 85 are generally considered within normal limits, with lower scores flagged as borderline or abnormal [2]. Always check the specific product’s scale rather than assuming a universal cutoff.

Should I be concerned about one elevated metric if everything else looks normal? Not necessarily. A single elevated metric, especially a borderline one, is weaker evidence than a convergent pattern across several metrics and other data sources. Consider the whole profile, the clinical history, and whether the elevation fits the presenting concern.

Can anxiety or poor sleep make a CPT report look abnormal? Yes. Confounds like anxiety, depression, and poor sleep can shift CPT performance without reflecting a stable attention problem. If a report doesn’t match the clinical picture, that’s a reason to look closer, not to automatically trust the number.

Do I need special training to interpret a CPT report? Basic interpretation of error rates, reaction time, and normative scores is described in each product’s manual and is within the scope of psychologists, psychiatrists, and PMHNPs who administer or order these tests. Complex or ambiguous profiles, or reports with validity flags, may warrant psychologist involvement or a second opinion.

Is a CPT report enough on its own to support a diagnosis? No. Every major review of CPT accuracy recommends against using a CPT in isolation. It’s one input alongside the clinical interview, developmental history, and standardized rating scales. See can a CPT diagnose ADHD? what it can and can’t show for the full answer to that question.

Start testing your patients with AxonCPT

A CPT report is only as useful as your ability to read it correctly and put it in context. AxonCPT’s report is built around the same principles covered here — normed against matched cohorts, broken into quarters so you can see the pattern across the session, and designed to be repeated at follow-up so you can track change. Start with 2 free tests every month, no card required. Create your free AxonCPT account or learn more at axoncpt.com.

References

1. Multi-Health Systems. Conners CPT Online Suite storefront and product brochures. Accessed September 2026.
2. The TOVA Company. How the T.O.V.A. Works and screening manual documentation. Accessed September 2026.
3. Arrondo G, Mulraney M, Iturmendi-Sabater I, et al. Systematic Review and Meta-Analysis: Clinical Utility of Continuous Performance Tests for the Identification of Attention-Deficit/Hyperactivity Disorder. J Am Acad Child Adolesc Psychiatry. 2024;63(2):154-171.
4. AxonCPT. axoncpt.com. Accessed September 2026.