A meta-analysis of memory ability in synaesthesia

People with synaesthesia (e.g., experiencing colours for letters and numbers) have been reported to possess enhanced memory relative to the general population. However, there are also inconsistencies in this literature and it is unclear whether this reflects sampling error (exacerbated by low Ns) or more meaningful differences that arise because synaesthesia relates to some aspects of memory more than others. To this end, a multi-level meta-analysis was conducted. Synaesthetes have enhanced long-term (episodic) memory with a medium population effect size ( $$  = 0.61), whereas the effects on working memory (short-term memory) were significantly smaller ( $$  = 0.36) but still exceeded that of controls. Moderation analyses suggested that, aside from the division between long-term vs. working memory, the effects of synaesthesia are pervasive, i.e., they extend to all kinds of stimuli, and extend to all kinds of test formats. This pattern is hard to reconcile with the view that synaesthetic experiences directly support memory ability: for instance, digit span (where synaesthesia could be helpful) showed a small effect whereas episodic memory for abstract images (where synaesthesia is irrelevant) yielded larger effects. Synaesthesia occupies a unique position of being the only known neurodevelopmental condition linked to a pervasive enhancement of long-term memory.