Electrophysiological Correlations of Third-Party Punishment

5034 words (20 pages) Dissertation

16th Dec 2019 Dissertation Reference this

Tags: Physiology

Disclaimer: This work has been submitted by a student. This is not an example of the work produced by our Dissertation Writing Service. You can view samples of our professional work here.

Any opinions, findings, conclusions or recommendations expressed in this material are those of the authors and do not necessarily reflect the views of NursingAnswers.net.

Chapter 1. Introduction

In these latter days, the problem of understanding human society have become more significant in different multidisciplinary fields of study.

Chapter 2. Social punishment

2.2 Social punishment and its types

Many forms of social mechanisms constitute human society. Social values, social institutions, control, communication are just a little part of them. However, what makes human to be human? What differs our society from the animal world?

The ability to develop and enforce social norms is probably one of the distinguishing characteristics of the human species (Fehr & Fischbacher, 2004a). These are informal rules that support maintaining normative standards of behavior. The process of internalization of social norms takes place next to socialization in early childhood and occurs during the entire life. The main mechanisms of enforcing social norms and stabilizing the behavior are reward and punishment (Bendor & Swistak, 2001)

Social punishment as an action strengthens compliance with the norm. Violation of the norms can lead to the feelings of anxiety, shame, guilt, an embarrassment that can be considered as first-party punishment (Elster, 1989).

Parties that are directly affected by the norm violation are so-called second parties. The case where the victims of norm disturbance retaliate against their own aggressors, this phenomenon is called “second-party punishment” (Figure 1, A). The desire to penalize the norm violator is rational because of the caused harm. However, there are social norms that can make harm nor directly to a person. In this case, the second-party punishment would not be a very effective way of teaching a lesson.

What is interesting and yet not good examined is another type of punishment that is called “third-party punishment” (Figure 1, B). Third party is an uninvolved party from outside of the situation of cheating or violation but happened to be a witness. An amusing matter of that is the fact that third parties in some circumstances show willingness to punish a violator even by incurring costs (Fehr & Fischbacher, 2004b).

Figure 1. A. Second-party punishment – victims retaliate against their own aggressors.
                               B. Third-party punishment – punishment of the norm violator from a third party.

Sanctions from the third party can come from one individual or from certain social group. Punishment can appear in a form of verbal (e.g. conviction, admonition), physical, financial (fine, deduction), social sanctions (exclusion, isolation). There are larger control systems even on a state level such as police and justice system (Marlowe et al., 2008).

To compare, second-party punishment strategies in iterated pairwise interactions are not evolutionarily stable in contradistinction to third-party punishment (Bendor & Swistak, 2001). In addition, third-party punishment is a good mechanism to minimize social norm disturbance and maintain the equality (Sun, Tan, Cheng, Chen, & Qu, 2015).

Another title for third-party punishment in so-called “costly” or “altruistic” punishment. This names demonstrate that this type of punishment can be an example of not rational choice, as there are costs that third party can be faces with. Reasons for such behavior could be explained by the motives of avoiding inequality and desire to reinforce cooperation within social groups (Fehr & Gächter, 2002). By this, we can suggest that third-party punishment is making an important contribution to the evolution of human altruism and this behavior was selected in human evolution (Fehr & Fischbacher, 2003). Altruistic punishment has no material benefit.

Evidence related to third-party punishment is coming mostly from ethnographic descriptions of social mechanisms in small-scale societies (Cronk, Changnon, & Irons, 1999). Another way of studying is exploring accounts provided by economic historians (Greif, 2013). Nevertheless, there is still not enough knowledge bound for this kind of punishment. However, there are still some was to gain knowledge of this phenomenon. The best way to acknowledge this evidence are provided in laboratory experiments mostly led by economic games.

2.3 Methods of studying social punishment

2.3.1 Dictator game

2.4 Three neuronal networks of third-party punishment

Frank Krueger and Morris Hoffman (2016) elucidated the role of neuronal networks in third-party punishment. By their study, in the process of blame and punishment are involved at least three huge brain networks: salience network (SN), default mode network (DMN) and central executive network (CEN) (see Figure 2) (Krueger & Hoffman, 2016).

Figure 2. Three neuronal networks of third-party punishment. SN (red): dAAC, AI, amygdala. DMN (blue): mPFC, PCC, TPJ, dmPFC. CEN (green): DLPFC, PPC.

The first network SN is located in the dorsal anterior cingulate cortex (dAAC) and anterior insula (AI). It also has vast connectivity with the amygdala. The main task that usually attribute to the SN functioning is processing averse experiences that will guide behavior (Bressler & Menon, 2010). By dAAC salience network find out whether there is any threat by the norm violation. Next, AI provides with aversive response to the information from dAAC. Finally, amygdala generates emotional response to the extent of the norm violation. Altogether, SN is furnished with the information of the degree of harm caused to the victim and emotional response that subsequently modulates the engagement of the next network. This network is an interesting object to study, as it is likely to be engaged in the mechanisms of all levels of punishment.

As it was mentioned above, salience network regulates the involvement of the second network – DMN that is anchored in the medial prefrontal cortex (mPFC). Default mode network in general is understood with the processes of autobiographical memory, mentalizing and self-monitoring (Bressler & Menon, 2010). In the processes of social punishment, DMN is associated with integration of the antecedent information from SN to evaluation of blame. There are also intranetwork connections with posterior cingulate cortex (PCC) and temporoparietal junction (TPJ). This area evaluates the intent of a wrongdoer or its goals and aim. This stage has an important role to play in punishment of second party mostly.

Finally, for making a certain decision to punish, central executive network (CEN) takes here a place. Generally, CEN is well known as context-dependent assessments for decision making in higher-order cognition. While posterior parietal cortex (PPC) is most likely to be responsible for constructing a scale of punishment, dorsolateral prefrontal cortex (dlPFC) selects specific way of punishing the norm violator within the context of crime (Buckholtz, Asplund, Dux, Zald, & Gore, 2008).

2.5  Event-related potentials of third-party punishment and salient network

Lu Sun and Peishan Tan (2015) investigated the effect of altruistic tendency on fairness in third-party punishment. There exploration was based on event-related potential (ERP) method using dictator game paradigm of third-party punishment (Sun et al., 2015).

At the basics of a neurological process of fairness consideration lie down two fundamental event-related potentials: Medial frontal negativity (MFN) and P300.

MFN is a negative event-related potential at frontocentral recording sites with peak amplitude between 200 and 350 ms. The source of this component is expected to be in the medial-frontal cerebral regions in particular the anterior cingulate cortex (ACC) (Fukushima & Hiraki, 2006). It was suggested that MFN in order to regulate behavior and take away the corresponding responses, uses information about the reward or non-reward (punishment) likewise from detector of the failure and success. This process is called “reinforcement learning theory” (Holroyd & Coles, 2002).

Main associations with MFN are performance evaluation such as feedback-related and error-related negativity. Another study shows the contribution of medial frontal negativity in social expectancy or norm violation. Research of the Sun (2015) showed that MFN varies in participants with different levels of altruism caused by different social expectancy.

P300, another ERP component that is mostly interested for researches of emotion and attention. It has been shown that P300 is sensitive to subject’s evaluation of results and following experience of emotions and attention (Yeung & Sanfey, 2004). Previous results also show that larger P300 negative feedback was generated by positive feedback. Furthermore, another studies show that P300 event-related potential is being developed larger in the condition of reward compared to non-reward or punishment condition.

Interestingly, increasing of activity in ACC (as a part of a salient network) can be revealed while committing errors by subjects, evaluating expectations of outcomes. In addition, ACC reflects on losses as a part of social punishment (such as unfairness). MFN is expected to be generated in ACC; it is an error-related negativity that reflects neuronal activity of social expectancy or norm violation and is widely associated with obtaining negative feedback. MFN as well as ACC is widely associated with reflection of negative outcomes in the social domain like unfair appeal by other subjects (Boksem & Cremer, 2010).

Taken together we can make a conclusion, that observing MFN is also a part of observation of activity in ACC and salient network in a broad sense.

Based on this literature review, we decided to investigate electrophysiological correlates of third-party punishment and replicate the ERP’s study of Sun et al. (2015). We hypothesize that Medial Frontal Negativity and P300 event-related potentials arise when a participant sees an unfair distribution between dictator and recipient in contrast with fair distribution (Hypothesis 1). Another hypothesis is that there should exist a correlation between event-related potentials and behavioral results (Hypothesis 2). Based on our hypothesis, we will conduct a data analyses of EEG data collected upon experiment of Oksana Zinchenko.

2.6  Exploration of rTPJ and rDLPFC

2.6.1 Previous study

The experiment was conducted by O.Zinchenko in collaboration with V.Klucharev and A.Belianin in the Centre for Cognition and Decision Making (CDM). Study investigated connectivity between rDLPFC and rTPJ using Transcranial Direct Current Stimulation (tDCS) and EEG recording. Twenty right-handed subjects with normal or corrected to normal vision participated in the study. Three participants were excluded from the experiment, due to the absence of any performance in behavioral task; therefore, the final number of subjects was seventeen. Each subject participated in three sessions separated by 7 ± 2 days. Before the start of the game, offline tDCS was applied for 15 minute.  The third-party dictator game lasted approximately 30–40 min, the whole experiment took about 2,5 hours.

According to the legend, the dictator was given 40 monitory units (MU) to his budget, which he could distribute between him and the recipient. Recipient was able only to accept the distribution of the points. Subjects as a sanctioners were able to punish the dictator by using even monitory units from zero to 18 from their own budget (consisting of 20 MU).  These monitory units were multiplied by 2 and subtracted from the dictator’s budget. The first stage of the game design is the moment when participants see the budget of the dictator and their own (Figure 3). At the second stage, participants can observe the distribution of the points committed by the dictator.  Next, participant should notice the instruction “choose how many points will be deducted from Player 1’s budget” and make a decision at the third stage. Final, fourth stage, sanctioner can observe how his decision proceed of dictator’s budget and on restoring of the justice. At the stage II, participant can observe fair – 0/40, 5/35, 10/30, 15/25, 20/20 or unfair – 30/10, 35/5, 40/0 distributions made by the Player 1.

Figure 3. Third-party dictator game design.

Electrodes were located on F8 and CP6 electrode position of the EEG 10/20 system for rDLPFC and rTPJ correspondingly. Using method of tDCS, experimenters disrupt rTPJ and rDLPFC activity independently or jointly. The first study was conducted with simultaneous modulation of rDLPFC and rTPJ activity in three conditions: anodal rDLPFC and cathodal rTPJ, cathodal rDLPFC and anodal rTPJ, sham condition (low intensity tDCS). Study 2 was a control study were anodal stimulation of the rDLPFC and rTPJ was applied independently in three conditions: anodal tDCS of the rDLPFC, anodal tDCS of the rTPJ and sham.

As a result, synchronization and desynchronization effect of rTPJ and rDLPFC network was not achieved by tDCS (only marginal effect was found). Nevertheless, EEG analysis revealed AEC connectivity between signals of the sensors corresponding to the rTPJ and rDLPFC; significant effect was found by the stimulation of rTPJ.

Referring to the assumption from this paper and previous studies, where it was suggested that transcranial alternating current stimulation of the frontal and parietal areas could influence positively on coupling of brain rhythms (Polania, Marius, Opitz, Grueschow, & Ruff, 2015). We decides to conduct a series of experiments in order to find phase and frequency of connectivity of default mode network (TPJ) and central executive network (rDLPFC).

Chapter 3. EEG project

The experiment was conducted by O.Zinchenko in collaboration with V.Klucharev and A.Belianin in the Centre for Cognition and Decision Making (CDM). Study investigated connectivity between rDLPFC and rTPJ using Transcranial Direct Current Stimulation (tDCS) and EEG recording. Twenty right-handed subjects with normal or corrected to normal vision participated in the study. Three participants were excluded from the experiment, due to the absence of any performance in behavioral task; therefore, the final number of subjects was seventeen.

At the beginning of each session, participants were prepared to EEG recording with F8 and CP6 electrodes replaced with stimulation electrodes of Starstim system. The EEG on 64 channels was recorded in the resting state and during the whole game with tDCS sham condition. The typical procedure of increasing current over 30 s and switching off for the rest of the stimulation (15 min) was used for the sham condition. As suggested, it should not have any effect on neuronal activity (Zhao et al., 2017), therefore, the data from EEG recording and behavioral data were used for ERP analysis in my project.

Participants were instructed to play the dictator game with third-party punishment. Dictator game lasted approximately 30-40 min, while the whole experiment took about 2.5 hours.

3.2 Study design

Study design consists of data analysis of the experiment described above and conducting correlational analyses with behavioral data.

3.3 Data analysis

EEG data preprocessing and analysis was performed using BrainStorm software. For preprocessing, following steps were applied:

1. Removing DC offset

2. Applying band-pass filter [1-40 Hz]

3. Removal and mark bad channels

4. Select channel position map

5. Removal of muscle artifact

6. ICA analyses for eye movement artifacts

For the next ERP analyses EEG data was epoched between -200 to 1000 ms in the condition when participant observed fair or unfair distribution. Data were averaged between all events and participants for fair (20:20; 27 trials) and unfair (30:10, 35:5, 40:0; 50 trials). Difference wave was constructed by subtracting fair grand average from unfair grand average. From this picture we observed our event-related potentials response.

Following Sun’s study, we pooled the same main electrodes of interest F3, Fz and FCz for frontal, Pz for parietal and Cz electrodes.

To see the significance of the effect, in 200-230 ms time interval of the difference wave was used Permutation test with FDR correction in BrainStorm software.

For the correlation analyses, we extracted negative amplitude peaks from the difference wave of five main electrodes for each participant (using Brainstorm and MATLAB software). For behavioral data, we constructed ratio of the reaction to fair and unfair offers.


Monitory Units – an amount of points. MU (unfair) – an amount of points that subject could expend so that fairness will be restored in cases of unfair of extremely unfair distribution offers (30/10, 35/5, 40/0). MU (amb) – an amount of points that subject could expend so that fairness will be restored in cases of ambiguous distribution offers (25/15).

To analyze the interaction between assumed event-related potentials and behavioral data we set a Pearson correlational analysis in SPSS software.

3.4 Results

On the difference wave, we observed significant negative peak on frontal electrodes and smaller positive peak in parietal electrodes with longer latency (Figure 4).

Figure 4. Fair, unfair and difference wave of Fz, FCz, F3, Cz and Pz channel on [90, 290] ms time interval.

Permutation test also showed a statistically significant negative deviation at [200; 230] ms time interval in frontal electrodes and positive deviation at [220; 250] ms time interval for the parietal electrode (see Figure 5). To conclude, we replicated the ERP findings, but with shorter latency in general in comparison to Sun’s study. For this reason, we can state that the first negative peak refers to Medial Frontal Negativity. The second one we cannot relate to P300 because of its sorter latency.

Figure 5. Permutation test

In contrast to Sun’ study, we found a significant correlation between behavioral results and negative amplitude peaks of the difference wave for MFN on 200-230 ms interval.

Fz F3 Pz Cz FCz Behavioral Ratio
Fz Pearson correlation 1 ,944** ,458 ,912** ,964** -,656**
Sign. (two-taled) ,000 ,065 ,000 ,000 ,004
N 17 17 17 17 17 17
F3 Pearson correlation ,944** 1 ,468 ,879** ,901** -,633**
Sign. (two-taled) ,000 ,058 ,000 ,000 ,006
N 17 17 17 17 17 17
Pz Pearson correlation ,458 ,468 1 ,691** ,494* -,382
Sign. (two-taled) ,065 ,058 ,002 ,044 ,130
N 17 17 17 17 17 17
Cz Pearson correlation ,912** ,879** ,691** 1 ,917** -,637**
Sign. (two-taled) ,000 ,000 ,002 ,000 ,006
N 17 17 17 17 17 17
FCz Pearson correlation ,964** ,901** ,494* ,917** 1 -,590*
Sign. (two-taled) ,000 ,000 ,044 ,000 ,013
N 17 17 17 17 17 17


Pearson correlation -,656** -,633** -,382 -,637** -,590* 1
Sign. (two-taled) ,004 ,006 ,130 ,006 ,013
N 17 17 17 17 17 17

As amplitude peaks are negative numbers, we can conclude that there is negative interaction between MFN and behavioral performance: the more negative is brain response, the more third-party will invest in the punishment.

Chapter 4. tACS project

3.1 Pilot study

In order to detect the best method of rewarding participants after an experiment with 184 trials (in tACS study) we will conduct pilot study with two conditions: randomized reward and cumulative reward. Eight subjects will participate in an experiment with the first conditions, another eight subjects will participate in the same behavioral experiment with another reward condition (between-subject study). The purpose of the behavioral pilot study is to reveal the type of the reward after the game that maintains motivation of the subject the most.

3.3 Transcranial alternating current stimulation of rTPJ

In order to reveal the phase condition for DLPFC and TPJ connectivity, we should know the main frequency on which these two networks have communication. For this aim, another Master’s student Maria Nikonova will implement following tACS experiment.

Nineteen to twenty-one healthy right-handed subjects will be invited to participate in experiment. As in the previous studies, subjects will play dictator game with third-party punishment condition, except for the method of stimulation.

One electrode will be located at CP6 to stimulate rTPJ and the reference will be places at the shoulder (Figure 6). An online-stimulation with three-frequency bands of stimulation and sham condition will be implyied with intensity of 1 mA. Eight blocks of alpha, betta, theta and sham conditions will be распределены by 184 blocks (within subject experiment).

Figure 6. tACS stimulation of rTPJ (CP6 channel location) with the reference on the shoulder.

By the method of repeated-measured ANOVA with multiple comparisons data analyses, we expect to see the decline of the performance during TPJ stimulation.

In addition, we expect to find the most significant effect in one frequency and this frequency should be the main frequency of rTPJ and rDLPFC connectivity.

3.4 Transcranial alternating current stimulation of rTPJ and rDLPFC

3.4.1 Participants

Nineteen to twenty-one healthy right-handed subjects with normal or corrected to normal vision will participate in our study. Subjects will be recruited voluntarily.

3.4.2 Procedure

Before the session, participants will be informed that they are playing an economic game against other participants located in another room in the same building. Subjects will be also informed about the condition of the reward (based on pilot study). After electrodes are placed, online tACS will be applied. A debriefing after each session will reveal whether

3.4.3 Study design

Dictator game with third party will consist of 184 trials. All the pictures of participants in dictator game were selected and only the most neutral photos will be included (with average attractiveness). The experiment is programmed using E-Prime software.

In-phase and anti-phase stimulation protocols will be used for stimulation in the following conditions: sham, main frequency and control frequency (based on previous study) with 1 mA intensity. One electrode will be placed on F8 channel location (for rDLPFC stimulation) and another one on CP6 channel location (for rTPJ stimulation) (see Figure 7 below). For in-phase stimulation, reference will be places on the shoulder.

Figure 7. A.tACS in-phase protocol stimulation of rDLPFC (F8 channel location) and rTPJ (CP6 channel location) with the reference on the shoulder. B. tACS anti-phase protocol stimulation of rDLPFC  and rTPJ.

3.4.4 Data analysis

For the data analysis will be applied repeated-measured ANOVA with multiple comparisons (Bonferroni correction).

3.4.5 Expected results

Based on our hypothesis and study design, we expect to find an effect in the main frequency condition in in-phase or anti-phase protocol.

In anti-phase protocol stimulation, we expect to change the intensity of punishment using the stimulation on the main frequency. Based on previous studies, we expect to increase the intensity of the punishment.

Chapter 5. Conclusion

We discussed most sufficient literature and articles related to the problem that we wanted to study – electrophysiological correlated of third-party punishment. In order to investigate three main neuronal networks of third party punishment suggested by Frank Krueger and Morris Hoffman (2016) two projects were conducted.

In the first project, the main goal was to study ERP’s, MFN in particular of ACC and salience network. We analyzed the data of the previous experiment conducted by Oksana Zinchenko. Our results show MFN with shorter latency and correlation between behavioral data and MFN, which indicate special relationship: the higher in the MFN, the more the third party will invest to punish the violator of the norms.

The second project aimed to investigate another two networks and connectivity between rTPJ (DMN network) and rDLPFC (CEN network). Project will pass three steps: pilot study, tACS of rTPJ, tACS of rTPJ and rDLPFC. Pilot study should provide us with the behavioral results and the best way of rewarding condition for participants. By tACS of rTPJ we expect to find main frequency of connectivity between rTPJ and rDLPFC. In the second experiment we will use results of previous experiment and by tACS of rTPJ and rDLPFC with three conditions (main frequency, control frequency, sham) we expect to find the phase characteristic of connectivity between this two regions.

To conclude, our findings will reveal the communication of rTPJ and rDLPFC, will provide us with better understanding of underlying neuronal processes of third-party punishment.

Cite This Work

To export a reference to this article please select a referencing stye below:

Reference Copied to Clipboard.
Reference Copied to Clipboard.
Reference Copied to Clipboard.
Reference Copied to Clipboard.
Reference Copied to Clipboard.
Reference Copied to Clipboard.
Reference Copied to Clipboard.

Related Services

View all

DMCA / Removal Request

If you are the original writer of this dissertation and no longer wish to have your work published on the UKDiss.com website then please: