Neural responses in the lateral prefrontal cortex and hippocampus of macaques navigating virtual environments
Abstract
The lateral prefrontal cortex (LPFC) and the hippocampus (HPC) are two brain regions that receive inputs from association cortices, and are involved in different types of memory, short- and long-term respectively. In an attempt to discover how these regions encode visual information in a virtual environment task, I recorded neural activity from the right mid-posterior HPC of two male macaques (Macaca mulatta) and from the left LPFC of two other male macaques. Then they carried out a learning task in a virtual environment, using a joystick to navigate. I compared selectivity for specific views of task-relevant parts of the maze and found view-selective cells in both regions. Notably, during this, I was intrigued by the bursting behaviour of the HPC neurons. To investigate whether the HPC might be using a burst code, I compared task period decoding using a burst rate or a firing rate code in both regions in the same task. I found that while firing rate codes performed better in both regions, this difference was much smaller in the HPC. A contributing factor to bursting is the intrinsic property of spike frequency adaptation (SFA), and I was able to assess this property in LPFC neurons, comparing between a minimalist saccade task and our virtual environment task, as well as measuring the property directly using in vitro patch-clamp methods. I found that SFA was similar across tasks, but more pronounced in vivo than in vitro for both broad and narrow, putative pyramidal and parvalbumin-positive interneurons (which largely did not adapt in vitro). Summed up, I was able to acquire interpretable signals from the macaque brain during complex virtual navigation tasks that I was able to analyse at a variety of scales. From neurons that encode specific views, to differences in the responses that do that encoding, and finally adaptation response properties of different types of neurons.