Prefrontal mechanisms of behavioral flexibility, emotion regulation, and value updating
Rudebeck et al. 23 June 2013. Nature Neuroscience
-In an important study investigating the role(s) of the orbitofrontal cortex, Rudebeck et al. found differential effects on reversal learning and emotional regulation between localized excitotoxic and aspiration lesions of the region. They noted, using an errors to criterion measure, that while there was improvement in accomplishing the serial reversals across during a single session (9 in total), only those animals with aspiration lesions exhibited significantly impaired performance relative to controls and to excitotoxic lesioned animals. Moreover, those animals undergoing the excitotoxic lesions continued to respond appropriately to emotionally-associated stimuli by exhibiting an increase in latency to retrieve food rewards, unlike previously reported aspiration-lesioned animals. Finally, using an object reinforcer devaluation task (with selective satiation on one of two different food rewards), animals with both the aspiration lesion and excitotoxic lesion failed to modulate their behavior in terms of reward choice as a function of the satiation.
Overall, this paper very convincingly takes the specialness of reversal learning and its association with the orbitofrontal cortex away. The behavioral deficits induced by OFC removal are more appropriately interpreted, as the authors claim, as an inability to revalue objects in line with biological need or as a failure in “representing and updating specific outcome expectancies to guide decisions.” Previous studies claim a specific deficit in reversal learning to the OFC were more than likely, according to Rudebeck et al., to arise from damage to the uncinate fasciculus mediating important temporal-frontal interactions. One caveat, per usual in lesion studies, it is important to recognize that white matter tracts near the OFC are necessary for efficient reversal performance, their disruption may results in a fundamental dysregulation of a large number of circuits across the brain and may not be causal in the actual behavior under normal neural function.
Canceling Actions Involves a Race Between Basal Ganglia Pathways
Schmidt et al. 14 July 2013. Nature Neuroscience.
-Schmidt and colleagues tested behavioral inhibition in rats on task in which the rats initiated a trial by nose poking a center port, received a go signal (1 or 4 kHz tone) to enter either side ports on the right or left, and then received reward for proper performance. In 30% of trials, following the go cue, the rats received a stop signal (white noise) to remain in the central port. Recording with a tetrode and sampling for areas in the subthalamic nucleus (STN) and the substantia nigra pars reticulata (SNr), the authors found low latency, transient neural responses to the stop signal in the STN, independent of whether or not the animals actually failed inhibit his response to the go cue. In contrast, they also report a longer latency response for cells in the SNr, selective only for conditions in which the animal successfully inhibited their behavioral response to one of the side ports. These cells in the SNr which correctly distinguished Correct stop from Failed stop clustered anatomically in the sensorimotor core of the striatum, a subregion previously discovered to project the superior collicilus.
Furthermore, when organizing striatal cells according to direction selectivity, the authors note that while on correct stop, failed stop, and slow go trials, the dynamics of the firing rates for these direction selective neurons was equivalent when examined relative to time movement, but different when aligned at the Stop and Go cue. On Failed stop and fast go trials, activity ramped up rapidly following the Go cue and was already above baseline at the time of the Stop signal. On the other hand, on slow Go trials and correct stop, activity increased similarly only for the first 100 ms, as the other trial types, but failed to reach any appreciable increase of significance by the time of the Stop signal. These results, in the mind of the experimenters, are consistent with a race model for the implementation of the cue-motor response. Their interpretation of a failed stop trial and the lack of SNr response, therefore, is that the “early arrival of striatal GABAergic input, [shunted] away the effects of glutamatergic inputs from the STN.” Although these results are important, I tend to shy away in an admittedly biased way from overly simple interpretations. Although the data is certainly consistent with a potential race model, how many other models in the infinite parameter space of models can similarly produce outcomes described in this paper? To further this investigation, optogenetic manipulation of the STN should be able to induce a stop, independent of the stop signal, and have a variable influence on behavior as a function of the ramp up of this cue-motor response program.