Saliency of Instructive Stimulus Affects the Activation of the Anterior Insula

765 

Abstract Submission 

Yoshimi Ohgami¹, Yasunori Kotani¹, Hajime Kageyama², Nobukiyo Yoshida³, Hiroyuki Akai², Shigeru Kiryu⁴, Yusuke Inoue⁵

¹Tokyo Institute of Technology, Meguro, Tokyo, ²The University of Tokyo, Minato, Tokyo, ³Niigata University of Health and Welfare, Niigata, Niigata, ⁴International University of Health and Welfare, Narita, Chiba, ⁵Kitasato University, Sagamihara, Kanagawa

Yoshimi Ohgami  
Tokyo Institute of Technology
Meguro, Tokyo

Yasunori Kotani  
Tokyo Institute of Technology
Meguro, Tokyo

Hajime Kageyama  
The University of Tokyo
Minato, Tokyo

Nobukiyo Yoshida  
Niigata University of Health and Welfare
Niigata, Niigata

Hiroyuki Akai  
The University of Tokyo
Minato, Tokyo

Shigeru Kiryu  
International University of Health and Welfare
Narita, Chiba

Yusuke Inoue  
Kitasato University
Sagamihara, Kanagawa

The stimulus-preceding negativity (SPN) is an event-related potential that reflects anticipatory attention. Many studies have suggested that the anterior insula (aINS) could be a source of the SPN. One of the features of the SPN is that it can be recorded before a feedback stimulus conveying information about the correctness of the response, while it cannot be recorded before an instruction stimulus that conveys information about how to perform the task. A possible explanation for the lack of SPN before an instruction stimulus is the saliency theory (Kotani et al., 2017). This theory posits that instruction stimuli are less salient compared to feedback stimuli, and reduced saliency could lead to the absence of SPN. If the saliency theory is correct, the aINS that is a core hub of the salience network should not be activated before instruction stimuli, and the region should be activated when the saliency of instruction stimulus is enhanced. In the present fMRI study, we manipulated the saliency of the instruction stimulus by adding reward information (reward and no-reward conditions) to investigate if the aINS activation was affected by the saliency of instruction stimuli.

Participants were 26 healthy adults who performed a time estimation task. In the task, there were two trial types: an instruction (Inst) trial and a feedback (FB) trial. The Inst trial was always followed by the FB trial. In the Inst trial, the visual stimulus showed an instruction about the second that should be estimated in the following FB trial (3, 5, or 7 seconds) as well as the amount of possible reward for the following feedback trial (reward condition: 10, 100, or 1000; no-reward condition: 0 Japanese yen). In the subsequent FB trial, the FB stimulus was presented after the button press with time estimate to inform participants about the correctness of their time estimate (the button press was too early, correct, or too late) and the amount of reward. In the reward condition (RW), participants received the reward for a correct response, and they did not receive any reward in the no-reward condition (NR). The experiment was conducted inside a 3-Tesla magnetic resonance scanner, and four contrast images: (1) Inst/RW, (2) Inst/NR, (3) FB/RW, and (4) FB/NR were constructed at the group level. Beta values reflecting the level of activation in the left and right aINS were extracted from the four contrast images, and these values were subjected to a repeated-measures analysis of variance (ANOVA) with hemisphere (left/right), reward (reward/no-reward), and stimulus (instruction/feedback) as factors. Subjective ratings on the saliency of the instruction stimuli were also collected to estimate the saliency of Inst stimuli.

Significant activations in the left and right aINS were found in three of the four contrast images, except for the Inst/NR contrast. The ANOVA on beta values of the aINS revealed that the beta values in the left aINS were increased in the RW than in the NR for instruction and feedback stimuli. Regarding the right hemisphere dominance before the instruction stimulus, the beta value of the aINS showed right hemisphere dominance in the Inst/RW, while there was no hemisphere difference in the Inst/NR. The subjective rating on instruction stimuli showed that the saliency of Inst/RW was larger than that in the Inst/NR.

As we hypothesized, the aINS was not activated before instruction stimuli. On the other hand, the aINS was activated when the saliency of instruction stimulus was increased. Previous studies show that the salience network of the brain exhibits right hemisphere dominance. In the present study, beta values in the aINS in the Inst/RW also showed right hemisphere dominance. These results support the saliency theory of SPN, and suggest that the saliency of the stimuli could affect the activation of the aINS that is a source of the SPN.

Reward and Punishment ¹

Attention: Visual ²

Cognition

FUNCTIONAL MRI