The Section on Neurophysiology studies the frontal cortex and related parts of the brain. In this project, we have focused on the mechanisms of abstract response strategies and judgments about supramodal inputs, such as relative duration and distance, in the selection of future goals. Supramodal signals are those that transcend a given sensory modality, and they are an especially cognitive form of neural activity. In the past year, this line of research has focused on the computation of spatial and temporal signals in the frontal cortex, in particular the computation of durations and distances. Choosing goals based on abstract strategies (Genovesio et al., 2006, J. Neurosci.) and relative distances and durations have in common a key feature of advanced behavior: an adaptive decision cannot be based simply on a familiar stimulus and a response to that stimulus. According to traditional animal learning theory, previously experienced stimuli, responses to those stimuli, and the outcomes of those actions determine an animals behavior: if you know an animals history, you can predict its decisions. But how can animals decide what to do when they encounter a stimulus for the first time, or when they must make a judgment about aspects of a sensory stimulus that transcend a sensory modality, such as time? Timing judgments are key to nearly every aspect of daily life, and they require knowledge about structured event complexes of a highly flexible kind. Misunderstanding structured event complexes and the timing of events is a common characteristic of Alzheimers dementia and is also one of the unfortunate consequences of frontal lobe damage. We reported previously the existence of timing signals (Genovesio et al. 2006, J. Neurophysiol.), strategy signals (Genovesio et al., 2005), and goal signals (Genovesio et al., 2006, J. Neurosci.) in the prefrontal cortex. In the former study, however, the subjects made no temporal judgments, and we could not examine relative timing. In the past year, we recorded single-cell activity from the frontal cortex as subjects made such judgments. In both the duration task and a distance task, two visual stimuli of different color appeared, one after the other. These two stimuli were separated by a variable delay period. At the end of the second stimulus, a second delay period was followed by the simultaneous appearance of the two stimuli, and one of these was chosen as a future goal. In the duration task, stimuli persisted for 50-3200 ms; in the distance task, the stimuli each appeared for 1 s at different distances from the center of the screen, ranging from 8-48 mm. The correct goal was the stimulus that had been farther from screen center. We tested average neuronal discharge rates during the first 400 ms of delay period following the second stimulus, when the decision has presumably been made. In the periarcuate frontal cortex, we sampled 182 cells during the duration task and 312 cells during the distance task, including 90 during both tasks. In periprincipal prefrontal cortex, these samples were 98, 94, and 46 cells, respectively. Of the periarcuate cells, 21% had relative-duration effects, the same percentage had relative-distance effects, and just 7% had both. Of the periprincipal prefrontal cells, 24% had relative-duration effects, only 8% had relative-distance effects, and 11% had both. In addition, some cells showed a pronounced order effect, indicating which color came first, others encoded which stimulus had lasted longer, and some showed absolute-duration effects. These findings indicate that during comparative judgments, frontal cortex neurons contribute to decisions about relative magnitude in both the spatial and temporal domains. Because relative temporal coding was common in periprincipal prefrontal cortex, our findings highlight another role for this area beyond spatial working memory, following our previous work (Lebedev et al., 2004) showing that attention signal had been misinterpreted as memory signals in many neurophysiological studies. Finally, for cells tested in both tasks, the neurons that encode comparisons in one domain only (temporal or spatial) were more frequent (31% in the two areas, combined) than those reflecting both (8%). This project also continued the exploration of neuronal mechanisms involved in selecting future goals. In this part of the project, we studied future goal choice based on two abstract response strategies, one called repeat-stay and the other called change-shift. For the former, when a stimulus repeated from one trial to the next, the correct future goal was the same as the most recent previous goal; for the latter, when the stimulus changed from one trial to the next, the goal also needed to shift from the most recent previous goal to a different one. Using this task, we have previously shown (Genovesio et al., 2006, J. Neurosci.) that one population of prefrontal neurons encoded future goals (F cells) and a largely separate population encoded previous goals (P cells). To examine neuronal interactions, we computed joint peri-event time histograms (JPETHs) for neuron pairs recorded simultaneously from different electrodes. JPETHs are two-dimensional histograms that display joint discharge frequency in a series of time bins for two neurons, thus assessing effective connectivity as revealed by transient cross-correlations. Our database of 1,441 cells included 57 F-P pairs, 39 F-F pairs, and 30 P-P pairs. F-P and F-F pairs showed many significant positive correlations (21% and 28%, respectively), but P-P pairs had only one (3%), a statistically significant paucity. Negative correlations were also rare. At the population level, both F-P and F-F pairs increased their correlations after cue onset, with latencies of 275 ms and 258 ms, respectively. The correlations remained relatively stable until cue offset, after which they decreased rapidly in F-F pairs but not in F-P pairs, which showed stable correlations until feedback arrived. Significantly correlated F-F pairs usually shared the same goal preference, whereas F-P pairs did not. Further, both F-F and F-P pairs showed their highest correlations when monkeys had to reject the previous goal. These findings suggest that future goals are selected through interactions among F-P pairs, with one component (the P cell) retrospectively coding a previous goal and the other prospectively coding a future goal, and usually a different one. Later, recurrent connections among F-F pairs support goal maintenance.
| Genovesio, Aldo; Brasted, Peter J; Wise, Steven P (2006) Representation of future and previous spatial goals by separate neural populations in prefrontal cortex. J Neurosci 26:7305-16 |
| Genovesio, Aldo; Tsujimoto, Satoshi; Wise, Steven P (2006) Neuronal activity related to elapsed time in prefrontal cortex. J Neurophysiol 95:3281-5 |
