Gene expression changes in long-term memory unlikely to replicate in the long term

Abstract

Identifying the cellular effects of memory-forming experiences on neurons which enable subsequent memory recall is a fundamental aim of neuroscience. The search for the engram could benefit from single cell RNA sequencing, which can estimate the mRNA expression of all genes in large samples of individual brain cells from animals exposed to specific experiences. A recent study used spatial transcriptomics and single cell RNA-sequencing to identify "transcriptional signatures in subpopulations of neurons and astrocytes that were memory-specific and persisted for weeks"1. However, because the authors did not account for multiple statistical comparisons2 and instead used an "unadjusted" threshold for statistical significance, the reported findings are likely dominated by false positives3. Moreover, the statistical analysis treated individual cells as independent samples without accounting for correlations across cells derived from the same biological tissue sample. Reanalysis of the studys data using appropriate, widely accepted statistical procedures, identifies no significant differentially expressed genes. This suggests the data do not support the authors claim to have identified cell type-specific transcriptional signatures of memory in the mouse basolateral amygdala.