به نقل از مرکز تحقیقاتی مهندسی عصبی:
این سمینار چهارشنبه اول دی ماه 95 در تالار کنفرانس IPM نیاوران برگزار میشود:
Long-term High-capacity Memories of Object Values in the Primate Brain
Wednesday, DEC 21, 2016 (10:00 - 12:00)
Dr. Ali Ghazizadeh
Graduated from John Hopkins and Berkeley Universities
Post Doctoral Fellow at Laboratory of Sensorimotor Research, National Eye Institute (NIH)
Abstract
During our lifetime we are bombarded by a large number of visual stimuli that need to be memorized and discriminated based on their ecological relevance such as their reward value. The basal ganglia, especially their posterior part, play an important role in such visual discrimination (Hikosaka et al 2014). Beyond this basal ganglia circuitry, the neural mechanism involved in such discrimination is not well-known. We addressed this issue by doing whole brain functional magnetic resonance imaging (fMRI) in two rhesus monkeys. Before the scans, the monkeys repeatedly viewed many computer-generated fractals (n>80) which were consistently associated with a large reward (Good objects) or a small reward (Bad objects) for >10 days. During block design scans (>1 day after training), these objects were passively viewed in the periphery (left or right, 6deg) while the monkey kept central fixation with no contingent reward. Thus the differential coding of objects in passive viewing would represent acquired stable object values rather than immediate reward expectations. Scan results showed widespread object Goodness coding (Good > Bad) along the ventral bank of superior temporal sulcus (STS) particularly in areas V4, TEO, TEa as well as in orbitofrontal cortex (OFC) and lateral prefrontal cortex (LPFC). Weaker effects were also observed in early visual areas (V1-3) and lateral intraparietal area (LIP). Various subcortical areas such as basolateral amygdala, ventral claustrum also showed significant Goodness coding (cluster corrected, alpha<0.01). Concomitant with neuronal discrimination, we found strong behavioral discrimination outside the scanner measured as strong gaze bias toward Good objects during free viewing sessions. To test whether neural and behavioral discrimination is maintained in long-term memory, we retested the same objects after 8-12 months with no visual exposure (memory period). Importantly many of the same visual areas along the ventral bank of STS fully retained their Good object discrimination despite long memory interval. There was significant discrimination in areas such as LPFC but with reduced magnitude. Notably free viewing of objects following the memory period revealed persistent gaze bias toward Good objects despite absence of any encounter for many months. Our results reveal a robust and large capacity neural mechanism for discriminating and orienting toward valuable objects. It is not known how such diverse brain areas acquire and maintain Goodness coding. One possibility is that Goodness coding arises in basal ganglia first which in turn train cortical areas. This hypothesis remains to be tested.
During our lifetime we are bombarded by a large number of visual stimuli that need to be memorized and discriminated based on their ecological relevance such as their reward value. The basal ganglia, especially their posterior part, play an important role in such visual discrimination (Hikosaka et al 2014). Beyond this basal ganglia circuitry, the neural mechanism involved in such discrimination is not well-known. We addressed this issue by doing whole brain functional magnetic resonance imaging (fMRI) in two rhesus monkeys. Before the scans, the monkeys repeatedly viewed many computer-generated fractals (n>80) which were consistently associated with a large reward (Good objects) or a small reward (Bad objects) for >10 days. During block design scans (>1 day after training), these objects were passively viewed in the periphery (left or right, 6deg) while the monkey kept central fixation with no contingent reward. Thus the differential coding of objects in passive viewing would represent acquired stable object values rather than immediate reward expectations. Scan results showed widespread object Goodness coding (Good > Bad) along the ventral bank of superior temporal sulcus (STS) particularly in areas V4, TEO, TEa as well as in orbitofrontal cortex (OFC) and lateral prefrontal cortex (LPFC). Weaker effects were also observed in early visual areas (V1-3) and lateral intraparietal area (LIP). Various subcortical areas such as basolateral amygdala, ventral claustrum also showed significant Goodness coding (cluster corrected, alpha<0.01). Concomitant with neuronal discrimination, we found strong behavioral discrimination outside the scanner measured as strong gaze bias toward Good objects during free viewing sessions. To test whether neural and behavioral discrimination is maintained in long-term memory, we retested the same objects after 8-12 months with no visual exposure (memory period). Importantly many of the same visual areas along the ventral bank of STS fully retained their Good object discrimination despite long memory interval. There was significant discrimination in areas such as LPFC but with reduced magnitude. Notably free viewing of objects following the memory period revealed persistent gaze bias toward Good objects despite absence of any encounter for many months. Our results reveal a robust and large capacity neural mechanism for discriminating and orienting toward valuable objects. It is not known how such diverse brain areas acquire and maintain Goodness coding. One possibility is that Goodness coding arises in basal ganglia first which in turn train cortical areas. This hypothesis remains to be tested.
