Interactive effects of music and prefrontal cortex stimulation in modulating response inhibition

Influential hypotheses propose that alterations in emotional state influence decision processes and executive control of behavior. Both music and transcranial direct current stimulation (tDCS) of prefrontal cortex affect emotional state, however interactive effects of music and tDCS on executive functions remain unknown. Learning to inhibit inappropriate responses is an important aspect of executive control which is guided by assessing the decision outcomes such as errors. We found that high-tempo music, but not low-tempo music or low-level noise, significantly influenced learning and implementation of inhibitory control. In addition, a brief period of tDCS over prefrontal cortex specifically interacted with high-tempo music and altered its effects on executive functions. Measuring event-related autonomic and arousal response of participants indicated that exposure to task demands and practice led to a decline in arousal response to the decision outcome and high-tempo music enhanced such practice-related processes. However, tDCS specifically moderated the high-tempo music effect on the arousal response to errors and concomitantly restored learning and improvement in executive functions. Here, we show that tDCS and music interactively influence the learning and implementation of inhibitory control. Our findings indicate that alterations in the arousal-emotional response to the decision outcome might underlie these interactive effects.

A two-way ANOVA [Response-type (Stop-error/Stop-correct/Go-correct/Go-error, within-  For the main analyses, phasic EDA response was measured within a 3 second window 108 following feedback to the correct or erroneous responses (after-feedback EDA) and we found Another ANOVA [Response-type (Go-error/Go-correct) x tDCS x Pre-post x Music-122 type] was applied to the before-feedback EDA in Go trials. There was a significant main 123 effect of Pre-post (F(1,64) = 15.20; p = 0.0001) (Partial Eta Squared = 0.19) indicating that 124 before-feedback EDA significantly decreased by practice in Go trials. However, there was 125 no significant three-way interaction between tDCS, Pre-post and Music-type (F(2,64) = 126 0.69; p = 0.51) (Partial Eta Squared = 0.021) indicating that in contrast to the after-127 feedback EDA, the before-feedback EDA in Go trials was not modulated by music and 128 tDCS.

129
Results of a control study to examine subjective feelings/emotions induced by high-and 130 low-tempo music 131 All 14 participants completed the ranking for all the 28 songs. For each music factor such as 132 familiarity we conducted a separate Chi-Square test of independence to examine whether 133 there was an association between the music tempo (High/Low) and other subjective aspects 134 of music such as preference or familiarity. A contingency table was formed for each factor.

135
The null hypothesis to be tested is that the music tempo is independent of the other factor 136 (such as Familiarity); that is the music tempo is not associated with the subjective perception 137 of familiarity. 138 We first examined whether there was an association between the music tempo and 139 Preference factor. No significant association was found between the music tempo and 140 participants' Preference (X 2 0.05, 4 = 2.15; p = 0.71) ( Figure S3a). There was no significant 141 association between the music tempo and Familiarity (X 2 0.05, 4 = 1.16; p = 0.88) ( Figure S3b).

145
Previous studies 73-79 have shown that music tempo is associated with mood changes 146 and induces different levels of happiness/sadness and therefore we asked participants to rate 147 the songs based on their perception of happiness/sadness associated with each song 148 (sad/neutral/happy). As predicted, participants associated high-tempo and low-tempo music 149 with happy and sad categories, respectively, which appeared as a highly significant 150 association between the music tempo and perceived happiness/sadness of the songs (X 2 0.05, 2 = 151 29.47; p = 0.0001) ( Figure S3e). This indicates that participants paid attention to categorizing each song and our analytical approach was sensitive enough to dissociate subjective 153 perception of music.

154
In our main tDCS study participants performed the cognitive tasks in different music 155 156 conditions (high-tempo music, low-tempo music or background noise) and we found a 157 significant interaction between music condition and tDCS effect in modulating the cognitive 158 task performance and arousal response (EDA). Findings from our control test clearly indicate 159 that subjective perception of 'preference', 'familiarity', 'lyrics' and 'novelty' did not differ 160 between the high-and low-tempo music conditions. This confirms that the observed 161 interaction of tDCS and music condition was not arising from these non-specific aspects of 162 music. We also confirmed that our control test was sensitive enough to detect the well-known 163 effect of music tempo in inducing positive and negative mood states 73-79 . 164 We conclude that the significant interactions of music conditions (high-tempo/low-   Subjective impression provoked by high-and low-tempo songs in terms of 'Preference', 'Familiarity', 'Lyrics', 'Novelty' and 'Happiness/Sadness'. Vertical axis shows the incidence rate of each condition.