Responses to noisy periodic stimuli reveal properties of a neural predictor

Wilsaan M. Joiner, Mark Shelhamer

Research output: Contribution to journalArticlepeer-review

5 Scopus citations


In programming motor acts, the brain must consider both internal and external noise sources: inherent variation in sensory estimates and changes within the environment. An interesting question in motor control is how reliable responses can be programmed in the face of noise and how these two noise sources interact. We study this by investigating the generation of sequences of predictive saccades to visual targets. First, eight normal subjects tracked targets that alternated at a pacing frequency (0.9 Hz) that promoted predictive behavior, for 300 trials. When tracking this perfectly periodic stimulus, there was variability in the timing of the saccades (intersaccade intervals) that was distributed around the interval of the stimulus (556 ms). We used this inherent variability to set the timing of subsequent stimuli; subjects completed three additional sessions in which the variance of the stimulus timing (the interstimulus intervals) had the same (1.0 SD), less (0.5 SD), or more (2.0 SD) variability than the subject displayed when tracking the perfectly periodic stimulus. Despite changes in stimulus timing variability, variance of the response timing (intersaccade intervals) was equal to the variance of the stimulus plus "inherent variance" (response variance when tracking a perfectly periodic stimulus). Examining the correlations between saccade latency and interstimulus interval, this relationship is interpreted as a tradeoff between reliance on previous saccade performance (intertrial correlations) and reliance on the current stimulus.

Original languageEnglish (US)
Pages (from-to)2121-2126
Number of pages6
JournalJournal of neurophysiology
Issue number4
StatePublished - Oct 2006

ASJC Scopus subject areas

  • General Neuroscience
  • Physiology


Dive into the research topics of 'Responses to noisy periodic stimuli reveal properties of a neural predictor'. Together they form a unique fingerprint.

Cite this