A model of partial reference frame transforms through pooling of gain-modulated responses

Area of Science:

• Neuroscience
• Computational Neuroscience
• Systems Neuroscience

Background:

• Neural responses in the posterior parietal cortex (PPC) often use intermediate spatial reference frames, blending sensory inputs like vision and audition.
• These responses are frequently modulated by gaze direction, suggesting complex integration processes.

Purpose of the Study:

• To model how mixed-frame neural responses in the PPC are generated.
• To differentiate between distinct types of mixed-frame responses observed empirically.
• To investigate the underlying mechanisms of spatial reference frame transformations in the PPC.

Main Methods:

• Developed a computational model simulating pooled, gain-modulated responses, analogous to complex cells in visual cortex.
• Simulated two specific types of mixed-frame responses characterized by differential gaze shifts.
• Adapted existing reference-frame measures to distinguish between different response types.

Main Results:

• The model successfully generated mixed-frame responses exhibiting differential shifts with gaze in horizontal and vertical dimensions.
• The model also simulated responses shifting differentially based on gaze start and end points along a single dimension.
• Distinguished these from a third type of response with partial, equal shifts across gaze changes.

Conclusions:

• Mixed-frame responses in the PPC can be explained by pooling gain-modulated neural signals.
• Empirical data suggest multiple mechanisms contribute to these mixed-frame responses.
• Analyzing these responses offers insights into the local organization of presynaptic computations in the PPC.