Neuropsychological investigations of the prefrontal cortex have shown a clear role for both PFo and PFm in reward-guided behavior and emotion (Izquierdo and Murray 2004;Izquierdo et al., 2005). Damage in either area leads to altered emotion and social behavior as well as disrupted autonomic responses to valued foods. Building on this work, neurophysiologists have found a role for both PFo and PFm in the signaling and prediction of feedback consisting of rewards and punishments. Although PFo and PFm damage is associated with difficulty adjusting behavior to changes in feedback, there is little understanding of how neurons in PFo and PFm adapt to changes in feedback. In particular, no one understands the brain mechanisms that allow us to alter our choices based on changes in feedback. Because reward and punishment feedback is processed in the amygdala, PFo and PFm, we need to understand how the amygdala interacts with these cortical areas to process feedback and change maladaptive choices. Autonomic responses provides an objective measure of emotional responses to feedback. The amygdala, PFo and PFm are key sites for processing reward information and generating autonomic responses. All three areas are reciprocally interconnected and neuropsychological studies have revealed both cooperative and contrasting functions of the amygdala and prefrontal cortex in feedback-guided behavior. In addition, neurophysiology studies have shown that the activity of single neurons in amygdala, PFo and PFm correlates with the value of expected reward feedback. However, little is known about the physiological interactions of these areas with each other or how these interactions produce autonomic and other emotional responses. To address these issues the present project examined the contribution of the amygdala to autonomic responses and to neuronal activity in PFo and PFm. Three subjects performed a reward-guided task. The subjects had previously learned that stimuli were associated with different magnitudes of reward feedback. Given the choice between two stimuli, they needed to choose the one that produced the most reward. We recorded the activity of single neurons in PFo and PFm, as well as autonomic and behavioral responses, both before and after lesions of the amygdala. Bilateral lesions of the amygdala abolished the differences in stimulus-reward encoding between the PFo and PFm. In addition, the proportion of neurons that encoded the quantity of expected reward, but not received reward, was also decreased relative to preoperative levels. This decrease occurred mainly in the PFo, with the result that amygdala lesions also abolished this preoperative difference between the PFo and PFm. These data indicate that input from the amygdala is more important for encoding of expected reward values in the PFo than in the PFm. We have also started to characterize the way in which PFo and PFm neurons signal the size and probability of rewards when these change over time. Animals face probabilistic and volatile rewards constantly, yet the neural mechanisms underlying the processing of such rewards is not well understood. Visual cues informed subjects of the number of drops of upcoming juice. Subjects were well trained with a set of 8 visual cues, including 2 fixed cues that predicted fixed amounts of juice, 2 fixed probabilistic cues that predicted rewards at fixed probabilities, 2 volatile cues that predicted juice that would change its number of drops every 50 and 200 trials on average, respectively, and 2 volatile probabilistic cues indicating rewards that changed probabilities every 50 and 200 trials on average respectively. The subject reported whether they expected a small or large reward;they learned to make consistent responses for fixed cues and to follow the changes of reward value or probability for volatile cues. To examine the contribution of frontal cortex regions to processing of probabilistic and volatile rewards we recorded the activity of neurons in both PFo and PFm. We found that the activity of PFo neurons reflected both the expected reward and the actual reward that the subject received, but surprisingly only in the later part of a trial. In contrast, the activity of PFm neurons indicated the expected reward immediately after the onset of visual cues, and maintained such information throughout a trial. Although activity signaling probability and volatility was present maximally in the cue period in both PFo and PFm neurons, the signal was not as prominent as that for expected reward. This study suggests that PFo and PFm play different roles in processing probabilistic and volatile reward information. It has been shown that animals value certain types of visually acquired social information, including facial expressions of emotion, dominance cues, and even whether another animal is receiving reward. To extend our work on how the brain codes value, in addition to studying the ways neurons signal value of stimuli that have been associated with reward, described above, we have also studied signals related to the value of social stimuli. Specifically, we have examined amygdala contributions to face processing, as a window on social cognition. Several studies have been carried out with functional magnetic resonance imaging (fMRI) techniques. Subjects perform a passive viewing task in which they view images of faces, presented one at a time. Some faces have a neutral expression (neutral) whereas others have an emotional expression: threat, fear grimace, or lip smack. One phenomenon reported in the fMRI literature is that facial expressions of emotion, relative to neutral faces, lead to greater activation (greater blood oxygen levels) in both the amygdala and in the sensory cortex dedicated to processing of visual information, namely, the inferior temporal cortex (IT). Because the amygdala is known to be important for processing the emotional content of images, and is anatomically related to IT, we hypothesized that the amygdala provided feedback to IT that resulted in the greater activation for emotional relative to neutral faces. To test this idea, we examined the effect of amygdala lesions on the pattern of activation in IT in response to viewing faces. Specifically, we studied three subjects with excitotoxic amygdala lesions and three controls using fMRI. Images of four distinct facial expressions were presented to the subjects in a blocked design: neutral, aggressive (open mouth threat), fearful (fear grimace) and appeasing (lip smack). Our results showed that in IT cortex of subjects with amygdala lesions the typical pattern of stronger activation to emotional relative to neutral faces was disrupted. By contrast, face responsivity (neutral faces >scrambled faces) and face selectivity (neutral faces >non-face objects) were unaffected. Overall, our data demonstrate that the feedback projections from the amygdala to IT cortex mediate the effect found there. Given the functional relationships between the amygdala, IT and frontal cortex, we investigated whether PFo or ventrolateral prefrontal (PFv) cortex might also contribute to valence-based response modulation, as described above. Both PFo and PFv are reciprocally connected with both the amygdala and IT, and could therefore modulate IT responses via direct projections, or via indirect projections involving the amygdala. We compared fMRI responses to facial expressions of emotion using an event-related fMRI paradigm. Cortical areas PFo and PFv showed higher responses for the expressive faces in two subjects. These results provide support for the idea that PFo and PFv contribute to valence effects seen in IT. The relative contributions of the amygdala and areas PFo and PFv to valence effects in IT deserves further investigation.

Project Start
Project End
Budget Start
Budget End
Support Year
6
Fiscal Year
2012
Total Cost
$1,408,769
Indirect Cost
Name
U.S. National Institute of Mental Health
Department
Type
DUNS #
City
State
Country
Zip Code
Mitz, Andrew R; Philyaw, Travis J; Boccuto, Luigi et al. (2018) Identification of 22q13 genes most likely to contribute to Phelan McDermid syndrome. Eur J Hum Genet 26:293-302
Rudebeck, Peter H; Ripple, Joshua A; Mitz, Andrew R et al. (2017) Amygdala Contributions to Stimulus-Reward Encoding in the Macaque Medial and Orbital Frontal Cortex during Learning. J Neurosci 37:2186-2202
Mitz, Andrew R; Chacko, Ravi V; Putnam, Philip T et al. (2017) Using pupil size and heart rate to infer affective states during behavioral neurophysiology and neuropsychology experiments. J Neurosci Methods 279:1-12
Kaskan, P M; Costa, V D; Eaton, H P et al. (2016) Learned Value Shapes Responses to Objects in Frontal and Ventral Stream Networks in Macaque Monkeys. Cereb Cortex :
Rudebeck, Peter H; Murray, Elisabeth A (2014) The orbitofrontal oracle: cortical mechanisms for the prediction and evaluation of specific behavioral outcomes. Neuron 84:1143-56
Rudebeck, Peter H; Putnam, Philip T; Daniels, Teresa E et al. (2014) A role for primate subgenual cingulate cortex in sustaining autonomic arousal. Proc Natl Acad Sci U S A 111:5391-6
Hadj-Bouziane, Fadila; Liu, Ning; Bell, Andrew H et al. (2012) Amygdala lesions disrupt modulation of functional MRI activity evoked by facial expression in the monkey inferior temporal cortex. Proc Natl Acad Sci U S A 109:E3640-8