Dissociation between the processing of humorous and monetary rewards in the ‘motivation’ and ‘hedonic’ brains

Humor elicits feelings of amusement and can be thought of as a social reward. We identified distinct mesolimbic reward system (MRS) processing patterns for monetary and humorous rewards. During both the reward anticipation and outcome phases, the nucleus accumbens (NAc) and anterior cingulate cortex (ACC) were active in response to monetary cues and monetary gains, while the amygdala and midbrain showed a differential response to humorous rewards, apparently driven by the hedonic enjoyment and appreciation of humor consumption. Psychophysiological interaction analysis (PPI) further demonstrated the functional coupling of the amygdala-midbrain circuit in response to humorous gains during the reward outcome phase, while neural signaling was observed in the NAc-ACC circuit during both the reward anticipation and outcome phases in response to monetary rewards. This is consistent with a view in which the NAc plays a key role in the ‘motivation brain’, and the amygdala in the ‘hedonic brain’. The findings further suggest that the neural mechanisms underlying reward consumption are more modality-specific than those underlying reward anticipation. Our study contributes to a growing understanding of neural responses to social rewards and represent an important first step toward understanding the neural processing of humor as one significant type of social reward.

2 objects, such as flowers, chairs, dogs, and houses. The humorous and neutral images used as stimuli were chosen by a group of five judges, all with experience in humor research. Each image was reviewed. Images with significant overlap or duplication of content were removed. An initial set of 600 humorous and 40 neutral images was selected. All the images were 16 × 16 cm, or with at least one side that was 16 cm long. The resolution of these images was at least 300 dpi, and they were presented on the screen at a proportion of 4 × 3. Examples of a humorous stimulus and a neutral stimulus are shown in Figure S1.  Table S1. A one-way repeated-measures ANOVA of the participants' funniness ratings was significant, F (1, 107) = 542.09, p < .001, ηp 2 = .835. 7 s and scores above 6 for comprehensibility and above 5 for funniness were selected.
The mean and standard deviation of the RT for comprehensibility was 4800.01 ± 1009.54 ms, of the comprehensibility ratings was 6.83 ± 0.11, and of the funniness ratings was 5.72 ± 0.21 (Table S2). The procedure is shown in Figure S2. The experiment consisted of the following three tasks: the monetary incentive delay task (MID), the revised humorous incentive delay task (HID), and the no reward task. Each task consisted of 16 trials. All participants viewed and rated each trial for wanting and liking on a 7-point scale using E-Prime 2.0 software. The means and standard deviations of the rated levels of wanting and liking in all three conditions (16 trials each) are shown in Table S3.

N-back task
A general description of the n-back tasks is provided in the Experimental paradigm sub-section of the Methods section in the manuscript. Figure S3 provides a visual illustration to support that description. For the 0-back trials, participants needed to respond to the target that was identical to a pre-specified letter ("X"). For the 2back trials, participants needed to respond to the target that was identical to the letter that appeared two letters previously ( Figure S3).

N-back procedure:
The present study used only two tasks-a working memory task (2-back) and control task (0-back). Participants observed stimuli on a visual display and provided n-back task judgments with a response box (Lumina, LS-PAIR). For each trial, an nbask task instruction (0-back or 2-back cues) appeared, lasting for 2000 ms. The nback task was followed by a delay lasting for 1000 ms. Stimuli were presented centrally with 500 ms duration, followed by a 2000 ms interstimulus interval (ISI).
Participants needed to respond by pressing one of two buttons on a response box with their dominant right index for targets (4 times in a trial) or middle fingers for nontargets (8 times in a trial) after the letter disappeared (see Supplementary Figure S4).
Each n-back trial lasted for 30 s (2500 ms × 12 letters). With the 2000-ms instruction 8 cue and 1000-ms retention delay, the total duration of a given trial is 33 s.
During the fMRI acquisition, participants were required to perform four separate runs. Each run consisted of 12 n-back trials. Each run was composed of two 0-back trials and two 2-back trials of each condition (monetary, humorous, and no reward conditions).

Figure S4
Experimental paradigm of N-back task. Each trial consists of an instruction cue, a retention delay, and an n-back task.

Analysis of n-back task (A) Percentage of accurate responses
The highest possible score for a given trial was 12. A correct response was given 1 point, whereas incorrect responses did not receive any points. There were 48 trials in the experiment, so the highest possible score was 576 (12 × 48 trials). The total scores participants received were converted into percentages of accurate responses.
(B) Two Criteria: successful or unsuccessful trials.
As the current study focused on the performance of reward outcomes, we set the following two criteria: (1) the total score of the responses in each trial should be greater than or equal to 9, and (2) the total score of the responses for targets in each trial should be greater than or equal to 2. If both of the criteria were met, participants saw the reward pictures, and the trial was counted as a successfully rewarded trial. We then calculated the total number of successfully rewarded trials. The total number of trials that failed to receive rewards was calculated by subtracting the total number of successfully rewarded trials from the 48 total trials in the experiment for each participant (Table S4).

Results of n-back task (A) Percentage of accurate of responses and two criteria
The percentage of accurate responses for the 0-back tasks was higher than for the 2-back tasks (see the upper part of Table S4). By applying the two criteria (75% and 50%), the pattern was reversed (see the bottom part of Table S4). This is because only 6 out of 38 participants reached at least 70% of accurate responses in the 0-back task in some given trials. However, they did not meet our two criteria, and yielded a smaller number of successfully rewarded trials in the 0-back task.
Only participants with more than 12 successfully rewarded trials (out of 16 trials in the experiment), and with the number of successfully rewarded trials larger than 4 in both the 0-back and 2-back tasks trials (out of 8 trials for both in each condition), were included in the present fMRI study.

(B) Reaction times on responses to the n-back task
This study consisted of three conditions, a monetary condition, a humorous condition, and a no reward condition. Each condition included eight 0-back trials and eight 2-back trials, for a total of 16 trials per condition. Each trial consisted of the presentation of 12 letters as stimuli. Each letter was presented in the center of the screen for 500 ms. After the 500 ms, the letter disappeared, and participants had 2000 ms to respond with their judgment. Participants responded by pressing one of two buttons on a response box with their (dominant) right hand to indicate whether the letter was a target or a non-target letter. The reaction time of judgment responses was recorded. Within each condition, participants thus made 96 judgments for each of the two n-back types (12 letters per trial, eight trials for each task type per condition).
When participants failed to respond, a reaction time of 2000 ms was recorded. The reaction times were thus calculated using 96 judgment responses (whether successful or unsuccessful) for both 0-back and 2-back tasks within each of the three conditions.
A summary of the data on reaction times is provided in Table S4. In summary, the percentage of accurate responses for the 0-back tasks was higher than for the 2-back tasks, but by applying the two criteria (75% and 50%), the pattern was reversed. Reaction times on the successful 0-back tasks in each of the three conditions were significantly shorter than reaction times on successful 2-back tasks. However, reaction times for unsuccessful 0-back tasks in the no reward condition were significantly longer than those for unsuccessful 2-back tasks in all three conditions (monetary, humorous and no reward).
According to the behavioral and fMRI results, the percentage of accurate responses was higher for the 2-back than for the 0-back tasks after applying the two criteria. In the post-scan debriefing, some participants reported feeling that the 2-back tasks were more challenging and generated higher motivation to earn the rewards in each condition, especially in the monetary reward condition. In contrast, the 0-back Note: 304 trials in total on each n-back task (38 participants). Participants received feedback regarding the two criteria of accuracy of the n-back responses across 16 trials (8 trials for each 0-back and 2-back task condition) for each condition (monetary, humorous, and no reward).