Reorganization of neural circuit for fear recognition after anterior temporal lobectomy (selected for oral presentation).

Introduction: Several studies demonstrate the critical role on processing emotional stimuli of mesial temporal lobe structures, which are the common pathologic substrate of Temporal Lobe Epilepsy (TLE). We used fMRI to examine the reorganization of neural circuits underlying fear recognition after anterior temporal lobectomy in a group of TLE patients. Methods: Seven right-handed patients with a history of drug-resistant TLE (three with right and four with left TLE), were evaluated before and six months after lobectomy. Six right-handed healthy volunteers were tested as controls and re-tested after six months. Subjects were asked to make a gender decision task on fearful (F) and neutral faces (N). In a control condition (C) subjects were asked to detect a white square within scrambled faces. EPI data were acquired using a GE Signa HHS77 system at 1.5 Tesla (TR=3380 ms; TE=40 ms) across 16 axial 5 mm slices (64 x 64 matrix) and were analysed using SPM99. Before scanning, patients underwent a neuropsychological evaluation to assess their facial emotion recognition abilities. Results: Controls: A conjunction analysis (F and N) showed increased signal in occipito-temporal regions and mesial temporal lobe structures bilaterally, consistently with the activation of a specific face-selective network. This pattern of responses was similar at re-testing. Fearful expressions evoked activations in the inferior frontal and posterior cortex (fusiform gyrus and temporal lobe) bilaterally; re-test data showed activations reduced in extent and restricted to the left hemisphere. Patients: Preoperative and postoperative data showed regions of increased signal for faces similar to those found in controls. The activation of distinct regions for processing fearful faces was present in five of the seven patients; it was missing only in the two patients with right TLE and early amygdala damage. These two patients also failed in explicit recognition of fearful expressions. After surgery, they improved their performance at neuropsychological testing, and their fMRI data showed activation areas partially resembling the fearful activations found in controls and in left preoperative TLE patients. In the other patients the pattern of responses to fearful faces was not consistent on re-testing. Conclusions: In control subjects the neural network activated by faces (either neutral or fearful) did not vary on re-testing. Neither TLE nor anterior temporal lobectomy affected the response of this network. On the contrary, right early amygdala damage impaired explicit recognition of fearful expressions and it was associated with lack of fMRI activations during incidental processing of fearful faces. Re-test data showed that the selective pattern of activation to fearful expressions varied with the re-presentation of the emotional faces in controls and according to the side and the nature of the preoperative damage in TLE patients. Indeed, anterior temporal lobectomy improved emotional recognition in patients with early right amygdala damage and was associated with activation of a neural network for incidental processing of fearful faces. We suggest that an early discharging right mesial temporal lobe damage can prevent the brain from undergoing functional reorganization. Right anterior lobectomy, removing the discharging tissue, can release brain plasticity mechanisms, leading to recovery of emotion recognition.