Spatial representation by ramping activity of neurons in the retrohippocampal cortex

Abstract

When solving navigational problems, remembering that a critical location is approaching can enable appropriate behavioural choices without waiting for sensory signals. Whereas multiple types of neuron in the hippocampus and retrohippocampal cortex represent locations using codes based on discrete spatial firing fields, analogue neural representations may be advantageous for efficiently recalling proximity to locations of behavioral importance. Here, we identify retrohippocampal neurons that use analogue ramp-like changes in firing rate to represent location as mice navigate a virtual environment in which they have learned the location of a reward. Ramp-like firing patterns had positive or negative slopes and could switch polarity or reset their rate at the reward location. These transitions were maintained when location cues were removed, indicating that path integration is sufficient to drive recall of the track structure expressed through ramp-like codes. We suggest that analogue ramp-like representations encode learned models for goal-directed navigation.

pyNeurode: a real-time neural signal processing framework

“…To evaluate the online spike sorting accuracy, we replayed a previously recorded 5 min 16-channel tetrode recording in the medial entorhinal cortex in Open Ephys [13], and let it pass through the entire data analysis pipeline, including filtering, whitening, spike detection in the Open Ephys GUI, and spike sorting in our framework. We then use the Spikeforest [4] package to sort the same recording with multiple offline sorters (Herding, Klusta, Ironclust and Spyking).…”

“…Our GUI is also running and displaying the real-time neuronal firing rate and spike waveforms. During the latency measurement, pre-recorded data from [13] are replayed in Open Ephys GUI. Therefore, the latency measurements also take the computing load of a normal running Open Ephys GUI into account.…”