Neuroscience Clinical Trial Endpoints Explained: What ADAS-Cog, CDR-SB, UPDRS, and PANSS Actually Measure

You read a neuroscience trial press release and the stock is up 40%. You don’t know why. The press release mentions ADAS-Cog, CDR-SB, and acronyms you’ve never heard of. Let’s fix that.

Neuroscience trials don’t measure tumor shrinkage like oncology. They measure how a patient thinks, moves, or functions using validated rating scales. Each disease has its own scale and its own language. Here are the ones that move stocks.

The problem

Oncology endpoints are concrete: tumors shrink or they don’t. Neuroscience endpoints are subjective: a clinician scores a patient on a scale, and that determines whether the trial succeeded. If you don’t know what the scales measure and what “meaningful” looks like, you can’t read a neuroscience press release.

Alzheimer’s disease: ADAS-Cog and CDR-SB

Alzheimer’s trials almost always use co-primary endpoints: one for cognition and one for function. The FDA requires both.

ADAS-Cog (Alzheimer’s Disease Assessment Scale, Cognitive Subscale): The standard cognition test. It measures memory, language, and praxis across 11 tasks. Scores range from 0 to 70, with higher scores meaning worse impairment. A clinically meaningful change is 2 to 3 points for patients with mild cognitive impairment (MCI) and 3 points for mild Alzheimer’s dementia. A drug that slows decline by 2 points on ADAS-Cog is doing something real.

CDR-SB (Clinical Dementia Rating, Sum of Boxes): The functional measure. It assesses memory, orientation, judgment, community affairs, home and hobbies, and personal care across six domains. Scores range from 0 to 18, with higher scores meaning worse function. A clinically meaningful change is 1 point for MCI and 2 points for mild Alzheimer’s dementia.

How to read the numbers: In the CLARITY-AD trial, Leqembi slowed decline by 0.45 points on CDR-SB at 18 months versus placebo. That’s a 27% slowing, but the absolute difference was less than the clinically meaningful threshold of 1 point for MCI. This tension between statistical significance and clinical meaningfulness is the core debate in Alzheimer’s drug development.

Red flag: A trial that hits ADAS-Cog but misses CDR-SB, or vice versa. The FDA wants both. Leqembi hit both, earning traditional approval.

Parkinson’s disease: UPDRS Part III

UPDRS (Unified Parkinson’s Disease Rating Scale) Part III: The motor exam. A clinician scores the patient on 33 items covering tremor, rigidity, bradykinesia (slowness), posture, and gait. Each item is scored 0 (normal) to 4 (severe), for a total range of 0 to 132. A clinically meaningful change is 2.5 to 5.2 points on the motor score.

The UPDRS also has non-motor sections: Part I (mentation, behavior, mood), Part II (activities of daily living), and Part IV (motor complications). But Part III is the one that makes or breaks Phase 3 trials. If the motor score doesn’t move, the drug probably doesn’t work.

What matters: Parkinson’s is progressive. A drug showing 3-point improvement on UPDRS Part III at 6 months that holds separation from placebo at 12 months is a winner. One that loses separation by 12 months is noise.

Schizophrenia: PANSS

PANSS (Positive and Negative Syndrome Scale): The standard scale for schizophrenia trials. It has 30 items, each scored 1 (absent) to 7 (extreme), for a total range of 30 to 210. Three subscales:

• Positive subscale (7 items): Hallucinations, delusions, conceptual disorganization, excitement, grandiosity, suspiciousness, hostility. These are the “psychotic” symptoms.
• Negative subscale (7 items): Blunted affect, emotional withdrawal, poor rapport, social withdrawal, difficulty in abstract thinking, lack of spontaneity, stereotyped thinking. These are the “deficit” symptoms, harder to treat.
• General psychopathology subscale (16 items): Anxiety, depression, motor retardation, unusual thought content, and others.

A clinically meaningful improvement is about 15 points (34% reduction from baseline). Most schizophrenia trials use the percentage of patients achieving at least a 30% reduction from baseline as the primary endpoint. If a drug shows a 10-point PANSS improvement, it’s statistically significant but may not be clinically meaningful.

Red flag: A drug that improves positive symptoms but not negative symptoms. Negative symptoms are harder to treat and matter more for long-term function. A drug that claims efficacy in “predominantly negative symptoms” is making an ambitious claim that requires a trial designed specifically for that endpoint, not a post-hoc subgroup analysis.

Multiple sclerosis: EDSS

EDSS (Expanded Disability Status Scale): The disability scale for MS. Scores range from 0 (normal neurological exam) to 10 (death due to MS) in half-point increments. A score of 4.0 means the patient can walk 500 meters without aid. A score of 6.0 means they need a cane to walk 100 meters. A score of 7.0 means they’re essentially in a wheelchair.

The EDSS changes slowly. A 1-point change can take years. This is why MS trials run 2-3 years and measure “time to confirmed disability progression” (a sustained increase in EDSS score of 1.0 point or more confirmed at a second visit). MS trials use EDSS as the gold standard, but newer endpoints like brain volume loss on MRI and NfL (neurofilament light chain) blood biomarkers are gaining ground as surrogate measures.

Red flag: EDSS scores of 6.0 and higher are less sensitive to change. A trial that enrolls only patients with EDSS 3.0-5.5 is designed to detect a treatment effect. A trial enrolling EDSS 6.0+ patients is harder to power effectively.

ALS: ALSFRS-R

ALSFRS-R (ALS Functional Rating Scale, Revised): A 48-point scale measuring 12 aspects of physical function: speech, salivation, swallowing, handwriting, cutting food, dressing, turning in bed, walking, climbing stairs, and three respiratory items. Each item is scored 0 (total loss of function) to 4 (normal), so the total ranges from 0 to 48. Lower scores mean worse function.

A clinically meaningful change is about 4 points (a 20% change). ALS patients typically decline 0.5-1.0 points per month without treatment. A drug that slows decline by 2-3 points over 6 months is doing something real. The FDA has also accepted survival as a co-primary endpoint in ALS trials: a drug that extends median survival by 2-3 months while showing a trend on ALSFRS-R.

Red flag: ALSFRS-R is sensitive to the rater. Inconsistent scoring across sites produces noisy data. Also, respiratory items (questions 10-12) carry disproportionate weight because respiratory failure is how most ALS patients die.

Why neuroscience trials fail more often

The Phase 2 to Phase 3 transition rate across therapeutic areas is about 30-45%. In neurology, it’s closer to 27%. Three reasons:

1. Placebo response is real. Patients in Parkinson’s and depression trials show significant improvement on placebo, especially in the first 12 weeks. A drug must beat not just baseline but a moving placebo target.
2. Disease progression is heterogeneous. Two ALS patients with the same ALSFRS-R score can progress at completely different rates. Trial randomization is the only defense, and small trials are vulnerable to imbalance.
3. Subjective endpoints. A clinician’s UPDRS Part III score depends on their training, their rapport with the patient, and even the time of day. Inter-rater variability adds noise that can wash out a real treatment effect.

How to read a neuroscience press release in 60 seconds

1. Find the primary endpoint. Is it ADAS-Cog, CDR-SB, UPDRS, PANSS, EDSS, or ALSFRS-R? If you don’t recognize the scale, the trial isn’t for your portfolio.
2. Check the p-value. p < 0.05 is the minimum bar. Above 0.05 and the trial failed regardless of what the press release says. Every trial’s primary endpoints are locked in on ClinicalTrials.gov before the first patient enrolls, so you can verify what the company committed to measure.
3. Check the effect size. A 2-point improvement on ADAS-Cog with p < 0.001 is statistically significant and probably clinically meaningful. A 0.3-point improvement on CDR-SB with p = 0.04 is significant and probably trivial.
4. For Alzheimer’s trials: both cognition and function endpoints must show benefit. If one hit and the other missed, traditional approval is off the table.
5. Check the placebo arm. If placebo patients improved by 3 points on UPDRS and the drug group improved by 6 points, the net treatment effect is 3 points. Report the net, not the absolute.

For the full framework on reading any clinical trial result, see how to read a clinical trial press release. For the statistical tools behind these numbers, see clinical trial readouts: hazard ratios, confidence intervals, and p-values.

Final checklist

• Do I know the disease-specific scale? (ADAS-Cog/CDR-SB for Alzheimer’s, UPDRS Part III for Parkinson’s, PANSS for schizophrenia, EDSS for MS, ALSFRS-R for ALS)
• Do I know what a clinically meaningful change looks like? (2-3 points on ADAS-Cog, 1-2 on CDR-SB, 2.5-5 on UPDRS, ~15 on PANSS, 4+ on ALSFRS-R)
• Did the trial hit its primary endpoint with a p-value under 0.05?
• Is the effect size clinically meaningful, not just statistically significant?
• For Alzheimer’s: did both ADAS-Cog and CDR-SB show benefit?
• Is the trial large enough? (300+ patients is signal; 30 patients is noise)

Neuroscience endpoints are not magic. They’re measurement tools with known strengths and weaknesses. Know the scale, know the threshold, and know the difference between a p-value and a result that changes how a patient lives.