1.

Introduction

Interest in bridging neuroscience and social psychology has seen a significant increase. Much of this interest has centred on brain localization—the attempt to relate social events to locations of brain activity. Social neuroscience, however, uses a rather unsocial environment.¹ Indeed, the brain activity of observers is often measured in isolation while viewing a variety of stimuli. The only social aspect is that employed stimuli involve other human beings.

This caveat of social neuroscience is mostly related to the technique employed, such as functional magnetic resonance imaging (fMRI), which suffers from several limitations. In this context, the most problematic limitation is that a proper and realistic social context is difficult to achieve. Due to this limitation, most previous studies have necessarily been conducted in an unusual and unrealistic way, such as using pictures shown on a computer screen. A functional brain imaging methodology that enables monitoring of brain activation in a more natural setting might well offer more informative data from more realistic social interactive situations, which is impossible with fMRI because of its methodological constraints.

In this study, we try to overtake these problems by measuring the cortical activity of individuals while they are engaged in a well-known social task, namely the social or interactive Simon, by means of near-infrared spectroscopy (NIRS).^2,3

The Simon effect is observed when people carry out spatially defined responses to non-spatial stimulus features, the location of which varies randomly. For instance, imagine that you are requested to press a left key in response to a green stimulus randomly presented on the left or right half of a display and a right key in response to a blue stimulus also randomly presented on the left or right half of the same display. Even though the stimulus location is irrelevant for the task, you will nevertheless be faster and more accurate if the green stimulus appears on the left side and the blue stimulus on the right one than if the green stimulus appears on the right or the blue stimulus on the left.⁴ That is spatial stimulus–response compatibility facilitates task performance. This phenomenon is known as the Simon effect. Usually, this effect disappears when participants respond to only one of the two stimuli, making the task a go/no-go task. Most accounts of the Simon effect are based on the assumption that it arises from a conflict between the spatial code of the stimulus and that of the response.^5,6 Accordingly, responses are assumed to be automatically activated if the stimulus corresponds spatially to the correct response, and thus it facilitates task performance, whereas a lack of correspondence between stimulus-response pair leads to response competition. At the theoretical level, the theory of event coding might account for this competition. This theory claims that perceptual representations (e.g., of things we can see) and motor representations (e.g., of hand actions) are linked, as the same representation (a common code) is shared by both perception and action.⁵ Thus, the simultaneous recruitment of two incompatible representations accounts for higher mean reaction time (RT) for incompatible trials as compared to compatible ones.

At the neuronal level, the Simon effect is thought to originate in posterior parietal structures,^7,8 which host attentional orienting and oculomotor mechanisms.⁹

This is not the whole story, though: the simultaneous recruitment of a common representational medium does occur not only when acting and perceiving in isolation but also when acting and perceiving in social context.^10–12 This claim receives its strongest support from studies on the social version of the classical Simon task. In this version, pairs of participants share a Simon task: one participant presses a key in response to one color, and the other participant presses another key in response to the other color, so that each participant actually performs a go/no-go task. This version of the task has been shown to produce a full-blown Simon effect: people perform better if the stimulus appears on the same side as the response key they operate. According to Sebanz and colleagues^10–12 this kind of social Simon effect suggests that each participant represents the stimulus-response rules and action plans of both agents involved. Thus, people may represent stimulus events irrespective of their intended target and may represent an action irrespective of who is carrying it out.

As regards the neuronal underpinnings of the social Simon effect, to date only two studies are available: one using EEG,¹² the other fMRI.¹³ In the EEG study, pairs of participants shared a Simon task: one participant pressed the left key in response to one color, and the other participant pressed the right key in response to the other color, so that each participant was performing a go/no-go task. Results showed that stimuli referring to the other’s action elicited an electrophysiological response, recorded from frontal electrodes, similar to that elicited by stimuli referring to one’s own action. Moreover, P300 component had higher amplitude in the group condition as compared to the single condition. According to the authors, these findings suggest that individuals acting in a social context form shared action representations. In the fMRI study, participants in the scanner performed the go/no-go task together with a confederate who was sitting next to the scanner and whose actions were visible in the lower part of a display. Results showed increased activation in ventral premotor cortex when participants acted upon stimuli referring to their own action alternative in the co-acting condition. Also, this condition was associated with increased orbitofrontal activation.

The above reported studies provide evidence that, while acting in social context, each participant represents the stimulus-response rules of both agents involved. Nevertheless, they do not help us to disentangle whether the social Simon effect originates at the level of parietal structures, as the standard Simon effect, or the level of sensorimotor cortex often activated during social tasks, likely via mirror mechanisms.^14–16 Indeed, both studies have intrinsic caveats, which are mostly related to the employed techniques. On the one hand, from the EEG study it is not possible to draw clear-cut conclusions about the cortical regions involved during the task. It is well known that EEG has a spatial resolution far from being accurate due to the distortion of the electric field while passing through the scalp.¹⁷ On the other hand, in the fMRI study, the participant being scanned could see the co-actor’s hand projected on the display, but they cannot surely share a common space, as it occurs in everyday social interaction, thus making the experimental context largely artificial and far away from realistic social interactions. This becomes particularly relevant if one considers Guagnano and colleagues’ report.¹⁸ They submitted participants to a social Simon task with a co-actor who sat either within or beyond their peripersonal space. Results showed that social Simon occurred provided that the co-actor sat within the participants’ peripersonal space. The authors claimed that the major role of the co-actor in the social Simon task might be to provide a spatial reference frame that allows coding of one’s own action as left or right relative to the other person—just as one’s own action alternatives provide a reference frame for relative response coding.¹⁹ A confederate sitting too far away (in the actor’s extrapersonal space) cannot provide the reference point for the spatial coding of the actor’s response.

The aim of the present study is to highlight the valuable contribution that NIRS can provide in studying social cognition. To this aim, we recorded cortical hemodynamic, from the left parietal and sensorimotor cortices, by means of NIRS, while 12 pairs of participants performed a social Simon task.

Two hypotheses were assumed: if the social Simon effect originates, as the standard Simon effect, at the level of attentional orienting and oculomotor mechanisms,^8,9,20 then higher activation should be found while processing no-go trials in parietal structures. On the other hand, if it originates from an actual mapping of the response of the other subject, then higher activation should be found in the sensorimotor cortex.

2.

Method

2.2.

Procedure

The experiment was conducted in a dimly lit room where participants sat side-by-side in front of a monitor (see Fig. 1). They were required to respond to digital photographs of a right hand with the index finger pointing to the left or to the right half of the screen. On the index finger there was a blue or green ring. Visual stimuli were presented on the central axes of the display, and the ring always appeared at the same location. Picture size was about $10\times 8$ visual degrees, in horizontal and vertical direction. The task was distributed between two individuals. Each person responded to only one of the two colors by pushing a single button with the right index finger, regardless of the pointing direction. For example, when one participant’s task was to respond to green, the co-actor responded to blue. This gave rise to four experimental conditions namely: (1) go compatible; (2) go incompatible; (3) no-go compatible; (4) no-go incompatible, where the go compatible trial refers to the condition in which the finger pointed toward the person who should respond; go incompatible trials refer to the condition in which the finger pointed away from the person who should respond. The very same stimuli were coded as no-go compatible and no-go incompatible for the co-actor, respectively. That is, if the finger pointed away from the participant and it was not her turn to respond, this trial was coded as no-go compatible. On the other hand, if the finger pointed toward the participant and it was not her turn to respond, this trial was coded as no-go incompatible.

Fig. 1

Experimental setting in the social Simon task. Hemodynamic activity was recorded from one subject during the experiment. Nevertheless, both the participants of each pair wore the fitting cup with probes and detectors.

Each stimulus was presented for 1000 ms followed by a fixation cross lasting between 1000 and 1500 ms. Participants completed one block of 200 trials, 50 trials per condition, presented in pseudo-random order. The sequence of trials was the same for all subjects. Stimulus presentation and data collection were controlled by a custom software (developed by Gaspare Galati at the Department of Psychology, Sapienza Università di Roma, Italy²¹), implemented in MATLAB (The MathWorks Inc., Natick, Massachusetts) using Cogent 2000 (developed at FIL and ICN, UCL, London, United Kingdom), and Cogent Graphics (developed by John Romaya at the LON, Wellcome Department of Imaging Neuroscience, UCL, London, United Kingdom).

3.

Results

3.1.

Behavioral Results

Error rate was 0.4%. Error trials were excluded from behavioral and fNIRS data analyses. Mean RT of the correct responses was calculated for each condition; responses longer than two standard deviations from the individual mean were treated as outliers and not considered (4.6% of the data set).

To compare performance on compatible and incompatible trials, a paired-sample t-test was performed between these two conditions. The analysis yielded significant results ($t_{23}=-4.6$; $p<0.001$, see Fig. 2). RTs were faster on compatible [mean $(\mathrm{SD})=376$ (9.3) ms] than on incompatible [mean $(\mathrm{SD})=385$ (9.4) ms] trials. This result suggests a full-blown social Simon effect, which means that a representation of the action alternative not actually under one’s own control was activated as well as when it was under one’s own control.¹¹

Fig. 2

Mean RT in the social Simon task. Error bars represent standard errors.

3.2.

NIRS Results

3.2.1.

Effects of compatibility on go trials

The linear contrast go compatible versus go incompatible trials was run with the aim of revealing cortical regions sensitive to the stimulus-response compatibility while processing stimuli referring to one’s own actions. Results showed higher activation in the sensorimotor cortex [$p<0.05$, FDR corrected, see Fig. 3(a)]. The opposite contrast did not reveal any significant activation. Analysis on deoxy-hemoglobin did not reveal any significant group activation.

Fig. 3

Cortical activations (0.05 FDR corrected) contingent upon: (a) go compatible versus go incompatible trials; (b) no-go compatible versus no-go incompatible trials. Group activation data are rendered on the cortical surface of a “canonical” brain. The black line represents the recorded area.

3.2.2.

Effects of compatibility on no-go trials

The linear contrast no-go compatible versus no-go incompatible trials was run with the aim of revealing cortical regions sensitive to the stimulus-response compatibility while processing stimuli referring to others’ actions. Results showed higher activation in the left inferior parietal lobule [$p<0.05$, FDR corrected, see Fig. 3(b)]. The opposite contrast did not reveal any significant activation. Analysis on deoxy-hemoglobin did not reveal any significant group activation. We do not report contrasts between go and no go trials because, as we do not have a “solo” condition, the activations merely due to motor response are not subtracted out.

4.

Discussion

Although perception and action have been widely investigated on the assumption that they can be completely accounted for by focusing on single individuals, several cognitive neuroscientists, experimental and developmental psychologists, and philosophers have recently argued for the need to take a social perspective on perceptual, motor, and cognitive activities. Indeed, over the last few years, more and more theoretical and empirical papers have been devoted to finding out the neural and cognitive processes underpinning basic social phenomena such as joint attention^37–39 and joint action^40–46 in development as well as in everyday adult life.

Currently the most commonly used brain imaging technique is fMRI,⁴⁷ which monitors brain activity by measuring BOLD responses. A major problem in utilizing this brain imaging technique to study the neural correlates of social phenomena is that the actual social dimension cannot be reproduced. Also, huge machinery dimensions, disturbing noise, together with the horizontal and unnatural position of the participant during scans’ acquisition, constitute the most limiting factors in the use of the aforementioned technique. On the contrary NIRS, which measures tissue oxygenation, has several advantages compared to other imaging techniques. First, NIRS allows the participants to be almost totally free in their movements. Second, it enables brain activity measurement in a natural setting. Participants, indeed, can undergo NIRS examination in a comfortable position, without any noise. These characteristics of NIRS are considered to be particularly suitable for social interaction studies.

In this study we investigated whether and to what extent the sensorimotor cortex and the parietal lobe are involved during social actions.^48,49

Our study provides two main findings. First, higher activation was found in the sensorimotor cortex when participants acted upon compatible stimuli. Second, higher activation was found in the inferior parietal lobule when the co-participant acted upon compatible stimuli. As regards the first main finding, previous studies have shown a recruitment of sensorimotor and parietal cortices while performing a standard Simon task. For instance, Liu, Banich, Jacobson, and Tanabe⁷ mapped the neural substrates of the Simon task with fMRI and reported that was more activation in the non-corresponding relative to the corresponding condition in the left posterior parietal cortex. In the same vein Rusconi and colleagues⁸ investigated, by using repetitive transcranial magnetic stimulation (rTMS), the role of the posterior parietal cortex in the Simon effect. They found a causal involvement of this cortical region in the task, that is, interfering with the activity in this area vanished the Simon effect.

To us, the activation of sensorimotor structures in go trials and the activation of parietal structures in no-go trials is particularly striking as it helps us to shed light on the nature of the joint Simon effect. According to previous studies we might have expected higher activation in the ventral premotor cortex while processing others’ actions; nevertheless, this was not the case. A possible explanation for this different result can be addressed. One could explain the difference on the basis of the different tasks employed. Indeed, while in previous studies the two participants simultaneously performed the required action, in our experiment participants never acted simultaneously. Thus, while in previous studies the production and the observation of a motor response co-occurred, in our task they are temporally segregated. In any case, our data suggest that the joint Simon effect does not occur at the level of sensorimotor cortex where the other’s action is mapped; rather it seems to occur at the level of shared attentional orienting mechanisms, that is the parietal lobule. Our results are in line with a previous study by Egetemeir and colleagues.⁴⁸ They investigated brain activation during real-life joint-action tasks using functional near-infrared spectroscopy. Participants performed table-setting tasks, either alone or in cooperation with a partner. They found that joint action produced stronger activations, as compared to acting alone, in a number of areas including the inferior parietal lobule.⁴⁸ Similarly, Koehler and colleagues⁴⁹ measured brain activity while participants observed reaching, grasping, and displacing movements from different perspectives. They found that observation from an egocentric perspective led to a higher activation in the inferior parietal cortex than observation from an allocentric perspective.

Possibly one may argue that, as we do not have a condition in which participants performed the task alone to compare with the condition in which a pair acted together, our data should be interpreted with caution. Although we agree with this warning, based on previous evidence^12,13,46 and on our behavioral data, we believe that our activations are somehow related to a social phenomenon.

The existence of a mechanism allowing us to integrate our own and others’ experiences has been suggested by neurophysiological, behavioral, and imaging studies in both humans and nonhuman primates. For instance, single cell recordings from monkeys’ brains have shown that there are bimodal neurons located in the ventral intraparietal area (VIP) that respond not only to tactile or visual stimuli delivered within the peripersonal space of the monkey but also to visual stimuli presented within the peripersonal space of another individual facing it.⁵⁰ At the theoretical level, the integration of one’s own experience with those of con-specific can be construed in terms of a shared we-centric space⁵¹ conceived as providing “a powerful tool to detect and incorporate coherence, regularity, and predictability in the course of the interactions of the individual with the environment.”

We processed both oxy- and deoxy-Hb. No deoxy-Hb activation survived to the group analysis. Diverging results between oxy-HB and deoxy-Hb are often interpreted as a lack of statistical power for the deoxy-Hb parameter. However, fNIRS studies on emotional modulation of brain activation^52–54 reported more robust results for deoxy-Hb compared to oxy-Hb. Even though our experiment immersed the subjects in a real and ecological social context, thanks to the physical presence of the co-actor, the associated potential emotional effects (e.g., embarrassment) could impact equally on all the experimental conditions, thus being nulled by the contrast analysis.

Summing up, we investigated the neural bases of a social interaction involving spatial proximity of two co-actors. Using NIRS, we were able to arrange an experimental setting closer to everyday life than previously done, which allowed us to shed new light on the role played by sensorimotor cortices during social interaction.