Abstract
Identical stimuli can generate neuronal response with highly variable response amplitudes and timecourses. We ask whether variability in neuronal response (one measure of noise) depends on behavioral state. To address this question, we used intrinsic signal optical imaging to study the hemodynamic response area V4 in individual macaque monkeys in 4 behavioral states: Attention-In, Attention-Out, Passive Fixation, and Anesthetized (cf. Tanigawa et al NS abstract 2008). Between 70–100 trials were acquired for 4 sec at 4hz per condition. First-frame subtracted reflectance change (dR/R) timecourses from regions of interest overlying V4 color, luminance, and orientation domains were plotted. Using condition-specific amplitude-normalized response, the deviation between the dR/R and the mean dR/R for each time frame was calculated and averaged across all trials. Mean deviation values and associated standard error were compared across conditions. Domain-specific comparisons showed that neuronal variability from hemodynamic responses were dependent on behavioral state. The timecourse of deviation values showed different temporal patterns of response variability for the different behavioral conditions. In the Anesthetized state, deviation values climbed monotonically over the 4 sec acquisition period. In contrast, in the awake states deviation magnitudes increased initially and then plateaued. Moreover, deviation values for Attention-In were initially greater than Attention-Out conditions but plateaued at smaller values in the latter part of the 4 sec acquisition period. Attention-Out and Fixation were indistinguishable. Behaviorally related effects appeared similar across color, luminance, and orientation domains in V4. These behavior-related differences in hemodynamic response variability in macaque V4 suggest a lower level of neuronal noise in the alert than anesthetized state and furthermore that, in the alert animal, attention has a predictable effect on neuronal noise over the timecourse of the task.
EY11744, DA023002, Vanderbilt Discovery Grant.