Active Brain & Nervous System

Dynamic Necessity of Posterior Parietal Cortex During Statistical Context Learning versus Inference

Funder Wellcome Charity-funded
Recipient organisation University College London
Funding Not disclosed
Period Oct 2025 — Sept 2028

In plain English

AI plain-English summary

A mouse hears a sound and must decide which category it belongs to—and a brain region called the posterior parietal cortex (PPC) may be essential for learning that task, but not for performing it once learned. This matters because the PPC is known to be active during many tasks, yet experiments have produced conflicting results about whether it is actually necessary. The researchers hypothesise that the confusion stems from a hidden variable: whether the animal is still building its internal model of the task (learning) or has already built one and is simply applying it (inference). By tracking the same neurons across both phases using two-photon calcium imaging in mice, and then temporarily silencing the PPC with optogenetics at each phase, they can test this directly. This is fundamental science. It will clarify a basic principle of how the brain allocates its computational resources—whether certain regions are required only during model-building, not model-use. A deeper understanding of this distinction could eventually inform how we design training regimes for tasks that require rapid adaptation, such as in flight simulators or surgical training, but that is a distant application. The immediate impact is a clearer map of causal necessity in the cortex.

View original technical description
The posterior parietal cortex (PPC) exhibits robust task-related activity yet shows contradictory patterns of behavioural necessity across studies. We hypothesise that PPC is preferentially necessary during statistical context learning—when animals build or update internal models—but dispensable during inference from stable models. Using longitudinal two-photon calcium imaging with a mesoscope, we will track individual neurons across three computational phases: initial learning, stable inference, and context adaptation. VGAT- Cre×GtACR1-flox mice expressing soma-targeted jGCaMP8s will perform a sound categorisation task with manipulated stimulus distributions. Trial-by-trial learning rates will quantify computational state, with optogenetic disinhibition testing PPC necessity at each phase. We predict PPC inactivation will impair performance when the learning rate is high but not when it is minimal (inference), with flexible remapping of the same neurons from mixed- selectivity to stimulus-selective encoding. Excitatory-inhibitory circuit dynamics will be characterised, testing whether opponent-inhibition motifs strengthen during inference, creating stable, transferable solutions. This research will establish when PPC is causally required for behaviour, demonstrating that the distinction between context learning and inference—rather than task difficulty or sensory modality—determines PPC necessity. Keywords: posterior parietal cortex, statistical learning, two- photon imaging, optogenetics, excitatory-inhibitory balance, decision-making, context learning, neural remapping

View the original record at the funder ↗

Researchers

Serkan Shentyurk (EPMC Awardee)

Related Research

Grants with similar aims, by meaning.

Cortical circuits underlying visual decision-making behaviors in mice
PriorCircuit:Circuit mechanisms for computing and exploiting statistical structures in sensory decision making
Neural mechanisms of learning, planning, and decision-making
Neuronal computation underlying the generation of a transitive inference in mice
Remembering an environment: does sensory gating contribute to context representation in the retrosplenial cortex?

Original classification

PhD Studentship (Basic)

Plain English summaries and category classifications on this site are generated by AI and may not perfectly reflect the original research.