White matter microstructure and statistical learning in the auditory and visual modality

Mårtensson, J. 1 , Arciuli, J. 3 , von Koss Torkildsen , J. 2 , Szczepankiewicz, F. . 1 , van Westen, D. 1 , Nilsson, M. 1 & Lindgren, M. 1

1 Lund University
2 University of Oslo
3 The University of Sydney

There are substantial individual differences in the capacity for implicit statistical learning (SL). While we know little about the neural mechanisms that underlie these differences, some of this variance may be traced to individual differences in white matter density (Flöel et al. 2009; Frost et al. 2015). We investigated whether performance in auditory and visual SL tasks were related to white matter microstructure. Participants were 21 young adults. The SL tasks utilized the embedded triplet paradigm where a familiarization phase containing either a continuous stream of unfamiliar cartoon figures (visual SL) or musical tones (auditory SL) was followed by a surprise test. Learning was significantly above chance in both tasks. We employed diffusion tensor imaging to investigate whether white matter microstructure was related to ASL and VSL performance. A whole-brain search in major white matter connections using tract based spatial statistics revealed a correlation between VSL and white matter microstructure (Radial Diffusivity) in the forceps minor (bilaterally), corpus callosum and areas adjacent to the left striatum and left hippocampus. Our findings indicate that areas that are involved in learning and reward systems show differential white matter microstructure that is related to individual differences in statistical learning in our visual task.