Retinotopic coding bridges perception and memory in the brain, study finds

By Dr. Liji Thomas, MDReviewed by Benedette Cuffari, M.Sc.Jan 3 2024

A recent study published in Nature Neuroscience reports that retinotopic coding may determine how information from the retina is processed in the brain cortex.

Study: A retinotopic code structures the interaction between perception and memory systems. Image Credit: Lia Koltryina / Shutterstock.com

Introduction

How does sensory signaling interact with the brain neurons to produce memory? This has long been a vital point of interest despite research findings showing that the accuracy with which these memory- and sensory-linked representations occur may be explained by patterned operations.

Conventionally, it is assumed that there is no common coding between these two types of cortical areas that control perception and memory. Retinotopy describes how visual input from the retina is mapped to specific neurons, especially those related to vision.

This retinotopic structure has been proposed to be replaced by abstract amodal coding in mnemonic areas. The default mode network (DMN) is an example of how the highest memory structures at the cortical apex occur due to signal propagation through visual areas.  

To date, it remains unclear how these two types of information can interact meaningfully when based on different neural codes. Recent studies have indicated that the DMN and similar high-level cortical areas exhibit retinotopy in the form of population receptive fields (pRFs) evoked by visual cues, which produce inverted response amplitudes.

What did the study show?

The current study suggests that retinotopy is present at all levels of the cortical neural processing layers, thereby allowing perception and memory areas to link functionally with structured interactions. This is based on functional magnetic resonance imaging (fMRI) findings that map visual pRFs.

Moreover, negative pRFs were found to be present at high concentrations in mnemonic areas as compared to perceptual areas. Thus, the pRFs in mnemonic areas (PMA) may bridge the gap between perceptual areas (SPAs) and temporal memory areas responsible for spatial memory through the use of retinotopic coding as the common substrate for the interaction of these areas.

Similar locations in the visual field were represented by paired -pRFs in SPAs and PMAs. This could be partly due to signals originating in the SPAs and passed forward to the PMAs.

The amplitude of -pRF activity in the PMAs was inversely associated with that of the +pRFs in the SPAs during recall tasks of familiar visual scenes like the participants’ kitchens. This reflects the occurrence of bottom-up sensory and top-down internal memory tasks, further supporting the researchers’ hypothesis.

The results also indicate spatially specific inhibitory interactions of these two types of areas, which would arise if they were structured by retinotopic neural coding. The opposing dynamics are stronger between paired +/-pRFs mapped to the same region of spatio-visual signaling during both perception and memory tasks.

However, since both these tasks pertain to situations involving maximal opponent dynamics of visual and memory systems, the scientists tested activity levels during a task where both systems interact, such as the perception of a familiar scene. This resulted in the same spatially specific inhibitory association of pRF activity in these areas, thus suggesting that this is how these neurons typically interact in the natural setting.

The dynamic interaction between these areas is built upon a scaffold of retinotopic coding, which directs and determines the directionality and magnitude of interactions between recall and perception systems within the neural framework.

What are the implications?

Our findings challenge conventional views of brain organization…a shared retinotopic code between externally oriented (perceptual) and internally oriented (mnemonic) areas of the brain structures their mutual activity.”

These linked areas form a mutually inhibitory set of paired mnemonic or memory-linked and perceptual areas that display opposing responses in specific regions. These functionally associated areas show such responses both during perception, where information is carried to the brain for processing, and recall, when stored information is retrieved.

Even at high levels of brain processing, visual inputs from the environment are clearly mapped to specific areas through the inverted retinotopic code. This could improve the understanding of how the DMN works during externally directed tasks.

Further research is needed to specify the extent to which retinotopic coding guides these types of interactions between perceptual and mnemonic neural areas and whether they are restricted to scenes or extend to all areas in the cortex.