Skip to main content
Explore URMC




Presented at the The Society for Neuroscience Annual Conference 2014

We previously reported that neurons in lateral prefrontal cortex (LPFC) represent behaviorally relevant speed and direction of visual motion during all stages of memory guided motion comparison tasks (Hussar & Pasternak, JN 2012, 2013). Here, we examined how spatial location is represented in LPFC during an analogous memory guided comparison task involving locations of motion stimuli. We analyzed dlPFC spiking activity and power spectra of local field potentials (LFP) during a task in which monkeys compared locations of two moving random-dot stimuli, labeled as S1 and S2, separated by a memory delay.

Analysis of spiking activity revealed location selective responses during both S1 and S2 and these responses were stronger during S2, most likely reflecting additional demands imposed by the comparison with the remembered location of S1. These observations were mirrored by LFP power in the β band of frequencies (12-30 Hz). β-band power showed significant location selective suppression during both S1 and S2, with stronger effects during S2. During the delay, many individual neurons showed gradual anticipatory activity modulation leading to the S2 onset, a pattern also observed in the β-band power of LFPs. In contrast with this close relationship between β-band power and spiking activity, θ-band LFP power showed task-dependent modulations dissociated from neuronal activity. Indeed, location selective delay activity in single neurons was transient and unrelated to the selectivity recorded during S1, appearing in different neurons at different times, a pattern analogous to that observed during the motion comparison tasks. However, LFP power in the θ-band (5-8 Hz) showed a strong enhancement during the memory period with respect to baseline power, which was location selective and decayed by the end of the delay.

These results reveal a striking parallelism between the way individual neurons represent sensory and spatial information during comparison tasks, suggesting that spatial location shares with the parameters of visual motion a common, or analogous neural substrate for representing sensory information in LPFC during such tasks. This substrate might be different for evoked and mnemonic representations. Similar dynamics and location selectivity in β-band LFP and spiking activity suggest that evoked representations engage populations with clustered representations of sensory attributes, which oscillate at β frequencies. Mnemonic representations during the delay period, instead, may be supported by more distributed circuit-level representations linked to θ-band dynamics.