The 'testing effect': functional neuroimaging and connectivity differences after a week delay

Marin-Garcia, E. 1, 2, 3 , Mattfeld, A. T. . 1 , Candon, K. C. 1 & Gabrieli, J. D. E. . 1

1 Massachusetts Institute of Technology
2 Basque Center on Cognition, Brain and Language
3 Ikerbasque

Retrieval practice during encoding has a beneficial influence on memory performance compared with study alone, which is known as the 'testing effect'. The neurobiological bases of the 'testing effect' are not yet well understood. In the present experiment participants were instructed to learn Swahili-English word-pairs. During the encoding phase, outside of the scanner, participants were assigned randomly to either the 'study' group, which only studied the word-pairs, or to the 'test' group, which both studied and tested the word-pairs. After a week delay, all participants had the same final cued-recall test in the scanner. The 'test' group when compared to the 'study' group showed enhanced activation and connectivity for successfully remembered word-pairs throughout regions typically associated with motor and speech processing (putamen, supramarginal gyrus). This may reflect retrieval of a more contextualized or rich representation. In contrast, the 'study' group relative to the 'test' group showed enhanced activation and connectivity in regions typically identified with successful retrieval for correct versus forgotten trials (prefrontal cortex, insula). This may reveal greater reliance on default retrieval processes by the 'study' group. Our results suggest that the memory benefit from retrieval practice during encoding may be related with creation of an enriched representation.