Paradigm

Patients were recorded in an emotional picture-viewing task (Kuhn et al., 2005; Huebl et al., 2016) 2–5 days after the first stage of the surgery for electrode implantation and prior to the second operation to connect the electrode to the subcutaneous pulse generator. During the task, participants were seated in the MEG scanner with a displaying monitor in front of them. Pictures selected from the Chinese Affective Pictures System (CAPS) (Bai et al., 2005) were presented on the monitor in front of them. The emotional valence (1 = unpleasant ⇒ 5 = neutral ⇒ 9 = pleasant) and arousal (1 = calm ⇒ 9 = exciting) of the pictures were previously rated by healthy Chinese participants (Bai et al., 2005). The figures can be classified into three valence categories (neutral, positive, and negative) according to the average score on emotional valence. As low-level properties of the figures, such as contrast brightness and complexity of the figures, are not measured or reported in the CAPS, only very dark or bright pictures were excluded from the paradigm. In our paradigm, each experiment consisted of multiple blocks of 30 trials, with each block including 10 pictures of each valence category (neutral, positive, and negative) in randomized order. Each trial started with a white cross (‘+’) presented with a black background for 1 s, indicating the participants to get ready and pay attention, then a picture was presented in the center of the screen for 2 s. This was followed by a blank black screen presented for 3–4 s (randomized). The task was programed using PsychoPy (https://www.psychopy.org/) with the timeline of each individual trial shown in Figure 1A. The participants were reminded to pay attention to the pictures displayed on the monitor and instructed to try to experience the emotions the pictures conveyed. An additional neutral picture was presented randomly three times per block, upon which the patients were supposed to press a button to ensure constant attention during the paradigm. All participants completed 2–4 blocks of the paradigm and none of them missed any response to the additional figure, indicating that they kept focus and that their working memory required for the task is normal. Pictures displayed to different participants are overlapped but not exactly the same; the average valence and arousal values of the displayed pictures are as shown in Figure 1B. There were significant differences in the emotional valence scores, as well as in the arousal scores for the presented figures of the three emotional valence categories (one-way ANOVA followed by Bonferroni post hoc test, F_2,24 = 14642.02, p<0.0001 for the valence score, and F_2,24 = 2102.55, p<0.0001 for the arousal score). The positive figures have the highest valence scores and highest arousal scores; the negative figures have the lowest valence scores; whereas the neutral figures have the lowest arousal scores.