Purchase this article with an account.
Geoffrey M. Ghose, Pamela Walsh; Temporal kernels of motion perception are sharpened by training and attention. Journal of Vision 2005;5(8):236. doi: 10.1167/5.8.236.
Download citation file:
© ARVO (1962-2015); The Authors (2016-present)
The perceptual and physiological consequences of visual attention and training have primarily been studied using tasks which are spatially, but not temporally, demanding. To investigate the effects of attention and training on temporal integration, we used a temporally demanding motion detection task. Two monkeys were trained to detect a brief pulse of motion (67–117 ms) inserted into one of two peripherally located arrays of randomly moving Gabors. Attention was investigated by shifting the probability of pulse occurrence over the course of a trial between the two arrays and had clear behavioral effects: performance and reaction times improved when the motion pulse occurred at a likely time and place. To investigate the mechanisms underlying this effect, perceptual spatiotemporal kernels were constructed by averaging the random motion preceding false alarms. By comparing kernels constructed according to a global motion metric with those constructed according to a local motion metric, we determined that the animals were spatially integrating over the entire array. Consistent with this conclusion, reaction times closely corresponded to the latency of the global motion kernel. To evaluate the effect of attention, we then compared temporal kernels associated with false alarms made during a high probability epoch with kernels constructed from low probability epochs. As expected, attention increased the sensitivity of motion perception. However, the effect of attention was not a simple gain change: attention sharpened the temporal kernel so that it was better matched to the duration of the motion pulse. Furthermore, we found that this attentional effect increased with training as the subjects were able to detect briefer motion pulses. These results contrast with a classification image study showing attention increasing spatial sensitivity but not selectivity (Eckstein et al, 2002) and suggest that attention can have distinct effects on spatial and temporal integration.
This PDF is available to Subscribers Only