Data from: Adaptation to hand-tapping affects sensory processing of numerosity directly: evidence from reaction times and confidence
Cite this dataset
Maldonado Moscoso, Paula A.; Cicchini, Guido Marco; Arrighi, Roberto; Burr, David C. (2020). Data from: Adaptation to hand-tapping affects sensory processing of numerosity directly: evidence from reaction times and confidence [Dataset]. Dryad. https://doi.org/10.5061/dryad.95x69p8gh
Like most perceptual attributes, the perception of numerosity is susceptible to adaptation, both to prolonged viewing of spatial arrays and to repeated motor actions such as hand-tapping. However, the possibility has been raised that adaptation may reflect response biases rather than modification of sensory processing. To disentangle these two possibilities, we studied visual and motor adaptation of numerosity perception while measuring confidence and reaction-times. Both sensory and motor adaptation robustly distorted numerosity estimates, and these shifts in perceived numerosity were accompanied by similar shifts in confidence and reaction-time distributions. After adaptation, maximum uncertainty and slowest response-times occurred at the point of subjective (rather than physical) equality of the matching task, suggesting that adaptation acts directly on the sensory representation of numerosity, before the decisional processes. On the other hand, making reward response-contingent, which also caused robust shifts in the psychometric function, caused no significant shifts in confidence or reaction-time distributions. These results reinforce evidence for shared mechanisms that encode the quantity of both internally and externally generated events, and advance a useful general technique to test whether contextual effects like adaptation and serial dependence really affect sensory processing.
The dataset was collected through customized psychophysical tests to assess the effect of numerosity adaptation in terms of change of PSEs, the confidence of the response and reaction-time.
PSEs were estimated from best fitting cumulative gaussian to the data.
Confidence and reaction-time via gaussian fits to the distributions of either aggregate or single subjects data.
Variability in the estimete of all statistics were measured through bootstrap simulations. Boostrap was also exploited to run no-assumption paired t-tests to asses wheter the results in the different conditions were statistically different.