Examining the effects of transcranial direct current stimulation on human episodic memory with machine learning

Abstract

We aimed to replicate a published effect of transcranial direct-current stimulation (tDCS)-induced recognition enhancement over the human ventrolateral prefrontal cortex [1] and analyse the data with machine learning. We investigated effects over an adjacent region, the dorsolateral PFC. We found weak or absent effects over the VLPFC and DLPFC. We conducted machine learning studies to examine the effects of semantic and phonetic features on memorization, which revealed no effect of VLPFC tDCS on the original dataset or the current data. The highest contributing factor to memory performance was individual differences in memory not explained by word features, tDCS group, or sample size, while semantic, phonetic, and orthographic word characteristics did not contribute significantly. To our knowledge, this is the first tDCS study to investigate cognitive effects with machine learning, and future studies may benefit from studying physiological as well as cognitive effects with data-driven approaches and computational models. Author summaryNon-Invasive Brain Stimulation techniques (in our case, transcranial direct current stimulation) are widely used among neuroscientists to map cognitive processes, for example, memory, decision making, emotional processing. In many cases, the sample size is limited, or methods applied for the data analysis are questionable. Here we propose a replication study aiming at confirmation of memory enhancement as a result of the application of anodal transcranial direct stimulation over the ventrolateral prefrontal cortex. Additionally, the involvement of the dorsolateral prefrontal cortex into episodic memory was analyzed. We used a larger sample size and applied different data analysis methods, including machine learning techniques. Surprisingly, we did not replicate the results of the original study, and we found weak effects of memory impairment after stimulating the DLPFC. However, what is most remarkable is that we have observed no significant effect of tDCS involvement on memory performance. Machine learning methods revealed no effect of linguistic factors on tDCS effect for both the original study and the replication at the level of individual trials and participants. Our findings highlight the importance of considering individual differences and data on the level of a single trial; in our data, participants memory responses resembled guessing behavior when recognition performance was measured by AUROC and highlight the need for modifications in the memory test and the use of other performance measures.