What if variability is not the noise in ADHD, but part of the signal? 

One of the more puzzling things about ADHD is that ability can appear to change from one situation to another. A task that felt impossible yesterday gets finished quickly today. Someone concentrates for hours on one problem and struggles to remain with another for ten minutes. Performance changes with sleep, interest, urgency, novelty, environment and sometimes for reasons that are difficult to identify at all.

Research does not particularly like this kind of inconsistency. We usually collect repeated measurements and calculate an average. Variation around that average is often treated as something to control, reduce or explain away.

A new study in Scientific Reports asks whether, in ADHD, some of that variability may itself be worth measuring. Researchers recorded pupil responses from 50 children, including 28 with ADHD and 22 controls, while they completed a visuospatial working-memory task. Pupil size is interesting because it changes not only with light but also with cognitive effort and arousal, providing an indirect physiological window into processes involved in attention. Instead of looking only at the average pupil response, the researchers examined how much that response varied from one trial to another.

Children with ADHD who were assessed off stimulant medication showed approximately 23% greater residual inter-trial pupil variability than controls after accounting for memory load, age and IQ.

The statistical story needs some caution. The corresponding participant-level group difference was significant before correction for multiple comparisons but narrowly missed the threshold after false-discovery-rate correction, with q = .051. This is a small study, and pupil variability is nowhere near being an ADHD biomarker.

But the question behind the result is larger than this particular dataset. What if averages are hiding something important? Two people can produce the same mean reaction time while arriving there very differently. One may perform consistently across 50 trials. The other may alternate between extremely fast responses, lapses of attention and periods of good performance. Their averages may look remarkably similar. Their cognitive experience may not.

This has been a recurring problem in ADHD research. Intra-individual reaction-time variability has been studied for years because inconsistency itself appears to characterise some aspects of ADHD performance. Pupillometry gives researchers another way of looking at that instability.

The medication part of the study is interesting too. Within the ADHD group, stimulant medication was associated with reduced inter-trial pupil variability and a change in the average pupil response during the probe period.

That should not be interpreted simply as medication “normalising” the ADHD brain. Stimulants can directly influence physiological systems affecting the pupil, and the researchers could not completely separate pharmacological effects from cognitive or session-related effects.

What the study does offer is another reminder that the average may not always be the most informative number. This matters well beyond laboratory tasks. A person can be capable of doing something without being able to produce that performance reliably whenever it is required. From the outside, inconsistency can easily be interpreted as lack of effort: if you managed yesterday, why can't you do it today? But capacity and consistency are different questions.

ADHD research has spent decades asking whether performance is better or worse. Perhaps we should also keep asking how stable that performance is.

Reference

Yadav, S., Aadya, & Shukla, J. (2026). Trial-level pupil dynamics reflect attention-related variability and medication-associated modulation in ADHD. Scientific Reports.

https://doi.org/10.1038/s41598-026-68871-2

Related reading

Explore my ADHD research & neuroscience FAQ for more information about auditory processing, attention and the relationship between laboratory research and everyday functioning.

You may also be interested in When the hearing test is normal but listening is still difficult, which examines the distinction between hearing sensitivity and complex everyday listening.

For a closer look at electrophysiological measures of auditory processing, read ADHD and auditory processing: where does the difference actually begin?