Study shows experience reorganizes content-specific memory traces in macaques

A new study in macaques suggests that how memory traces are reorganized by experience is reflected not only in synaptic networks but also in the physiological “fingerprints” of the neurons that participate. The researchers report that key assembly-level features—such as which neurons belong exclusively to particular assemblies, how connected those assemblies are, and how sharp-wave ripples (SWRs) modulate network activity—can be shaped by differences in the cell types contributing to each assembly.

To probe this idea, the team analyzed neurons classified by their membership history: cells belonging exclusively to newly formed, recent, or older assemblies. Neurons with divided memberships were handled separately, ensuring that comparisons reflected genuine differences in assembly participation rather than mixed recruitment across time windows.

They then measured four physiological features that can differentiate inhibitory and excitatory neurons in vivo: global firing rate, burst index, waveform trough-to-peak duration (a measure related to spike shape and width), and interspike interval (ISI) distributions. This approach leverages known electrophysiological signatures—such as the expectation that inhibitory basket cells typically show higher firing rates, stronger spike narrowness, lower bursting, and shorter ISIs compared with pyramidal neurons.

Surprisingly, firing rates and burst indices did not significantly differ among the assembly groups, suggesting that overall activity level and burst propensity were not the main drivers of assembly-specific network behavior. Instead, spike morphology and timing carried the strongest signal. The spike width metric indicated narrower spikes for neurons in the recent and old groups relative to those in the new group, pointing to systematic changes in waveform characteristics as assemblies age.

In parallel, the ISI analysis revealed longer interspike intervals for neurons in the new group compared with those in recent or old assemblies. The authors report statistical support for this effect (permutation testing with false discovery rate correction; P < 0.01), reinforcing that temporal firing structure—not merely how often neurons fire—is reorganized during experience-dependent memory trace updating.

Together, these findings suggest a cell-type-sensitive mechanism underlying assembly exclusivity and connectivity dynamics. As experience reshapes memory content, the physiological regime of recruited neurons—especially spike width and ISI timing—may shift in a way that supports stable yet adaptable network representations.