Task Co-Representation in Aging: An Event-Related Potential Study

The goal of the present study was to investigate age-related changes in attentional allocation for shared task representations during joint performance; event-related potentials were recorded while participants performed a modified visual three-stimulus oddball task, both alone and together with another participant. Younger adults and older adults (14 each) participated in the study. Participants were required to identify rare target stimuli while ignoring frequent standards, as well as rare non-targets assigned to a partner’s action (i.e., no-go stimuli for one’s own task). ERP component, nogo-P3 and P3b were measured to investigate the inhibition and the attentional allocation to the partner’s stimuli. Results showed that younger adults elicited larger frontal no-go P3 and parietal P3b for non-targets in the joint than in the individual condition. Contrary to expectation, older adults induced frontal no-go P3 in the joint condition not in the individual condition. In the sharing of the task with another, the result suggested that the efficiency of matching of incoming information with the representation of the other’s task declined with age, whereas aging did not affect the suppression of incorrect preparation of motor responses instigated by this representation.

viduals vary with context and circumstances. This crucial ability has been attributed to a common coding system that links perceived events to intended actions and activates the same representations or common codes [5]. Neural evidence for common coding has been identified in mirror neurons in the premotor cortex of the macaque monkey [6] [7]. Common codes enable us to infer another's intention from observed actions, activating representations in our own action system. Shared representations with partners have been demonstrated using a joint action paradigm in a number of behavioral studies [8]- [13] as well as event-related potential (ERP) studies [14] [15] [16] [17] [18].
Typically, the joint action paradigm provides two conditions: a joint and an individual condition. A joint condition requires two partners sitting next to each other to respond to different types of information with the same fast response task embedded in it. In an individual condition, a participant, alone, is asked to respond only to the specified stimulus information, not to the other, i.e., go-nogo task. In our previous study [15], ERP activity was recorded while participants performed an auditory three-stimulus oddball task alone (individual condition) and with another participant (joint condition). In this paradigm, participants were required to identify frequent standard tones and rare target tones using designated keys, while ignoring rare non-targets assigned to a partner's action (i.e., no-go stimuli for the participant's own task performance). This research focused on the P3 component of ERP in a go-nogo task. In the go-nogo task, two types of P3 are observed [19]. For no-go trials, the no-go P3 is evoked in frontal and central sites, whereas go trials elicit P3b in central-parietal maximum. Parietal P3b has been measured as a "culmination of multiple cognitive processes" [20], such as attentional allocation [21] [22] [23], context updating [22] [24], and the timing of stimulus classification [25] [26], while frontal no-go P3 has been found with response inhibition or suppression of action tendencies [16] [17] [18]. Kato et al. (2016) observed P3b for targets in both individual and joint conditions, whereas P3b for non-targets was elicited only in the joint condition [15]. P3b for non-targets was interpreted as a reflection of intentional stimulus classification for non-targets, based on the notion of co-representations of one's own and another's actions formed in a joint setting. In addition, the emergence of P3b and no-go P3 for non-targets in the joint condition was delayed relative to P3b for targets, suggesting that shared task representations are K. Kato The present experiment was designed to examine age-related changes in attentional allocation for shared task representations during joint performance. A three-stimulus oddball paradigm was combined with a cyberball situation, in which participants played a ball-tossing game on a computer [27]. This combined paradigm was used to emphasize collaborative behavior in a joint action situation and to keep participants interested in performing the experimental task. ERP activity was recorded from each participant who performed the task alone beside an empty chair (individual condition) and from paired participants who sat side by side (joint condition). In these social contexts, three classes of stimuli (standard, target, and non-target) were presented on a computer monitor, which initially displayed four squares (boxes) in a square arrangement. A ball (black circle), the standard, was presented frequently and alternately in the upper left and right boxes, as if it were tossed between those boxes. For left-seated participants, the balls delivered to the lower left and right boxes were assigned as the target and non-target, respectively, and participants were required to press a left button for targets (i.e., go trials) and withhold button-pressing for non-targets and standards (i.e., no-go trials). For right-seated participants, the assignment of stimuli (targets and non-targets) and response-button position were reversed. ERP indices included frontal no-go P3 and parietal P3b.
We hypothesized that ERP differences in older adults would differ from those in younger adults. Specifically, we predicted that cognitive decline in older adults would make it difficult for them to allocate attentional resources appropriately for representation of the other's task in a joint context. This prediction would be revealed by non-target P3b. For younger adults, non-target P3b should be larger in the joint condition than in the individual condition, because younger adults, who have abundant attentional resources, can readily form a representation of the other's task and apply it to stimulus matching, as seen in the previous study [15]. On the other hand, for the above reasons, older adults, for whom attentional resources are less abundant relative to younger adults, should not show any differences between social contexts in terms of non-target P3b. In addition, we predicted that the processing of misleading stimuli would elicit frontal no-go P3, which reflects incorrect preparation and subsequent suppression of motor responses, for the participant's own task-irrelevant non-targets. On the other hand, for older adults, difficulties in forming and applying a partner's task representations should result in attenuation of no-go P3 measures for non-targets in the joint condition. P3b for targets should not differ between social contexts, because processing of stimulus classifications is executed based on one's own representations. In the individual condition, both younger and older participants should show a large P3b for targets, as in a three-stimulus oddball task [28] [29].
There should be no difference in P3b for frequently occurring standard stimuli Journal of Behavioral and Brain Science between the joint and individual conditions, as in the previous study [15].

Participants
Fourteen younger adults (age range: 20 -30 years, M = 21.8, SD = 2.8, 1 male and 13 females) and fourteen older adults (age range: 68 -74 years, M = 70.9, SD = 1.7, 8 males and 6 females) participated. The number of participants was determined by a previous study that investigated attentional control in aging using a three-stimulus oddball task [30]. Older adults were recruited from a community silver human resource center. There was no difference in years of Each pair of participants in the joint condition was acquaintances. Younger adults were selected from the same class in the same university department.
Older adults were members of the same community silver human resource center, who had participated several times in the authors' research and had previously met each other.

Stimuli
As shown in Figure 1, a computer monitor initially displayed four squares (boxes) in a square arrangement. All four squares were presented continuously during the block. There were three stimulus classes: standard, target, and non-target. The standard (appearance probability (p = 0.6) was a black circle  in diameter. All stimuli were displayed on a white background.

Procedure
The experiment was conducted in a quiet room. In the joint condition, paired participants sat side by side in front of a computer monitor, and ERPs were recorded from both participants. In the individual condition, each participant sat next to an empty chair. Viewing distance was 70 cm from the monitor. In both social contexts, participants were told to play a ball-tossing game. They were in-

Performance
Performance data are summarized in Table 1   The ERP components to be analyzed were selected on the basis of visual inspection of grand averages. P3b, which is a notable component for an oddball

400 -700 ms Time Window
In the 400 -700 ms time window, a significant main effect of age group was found for standards, F(1, 23

Discussion
The purpose of the research was to examine the effect of aging on the formation Journal of Behavioral and Brain Science of other's task representation in joint action. In this experiment, ERPs were recorded from younger and older adults performing a modified visual three-stimulus oddball task while alone (individual condition) and together with another participant (joint condition). Our previous study of younger adults [15] suggested that co-actors share task representations during stimulus matching. The purpose of the present study was to investigate whether cognitive decline in older adults makes it difficult for them to allocate attentional resources appropriately for representation of another's task.
First, for younger participants in the individual condition, our oddball procedure using visual stimuli replicated the robust findings of a large P3b for targets in a three-stimulus oddball task [28] [29]. By contrast, this large P3b for targets was not observed for older adults. This difference between younger and older adults may have emerged because older adults were less able to classify incoming stimuli unambiguously. According to the P300 context-updating model [32], as a stimulus enters the processing system, it is compared to the previous stimulus, and a sameness judgment is made. An enhancement of P300 amplitude is observed when the incoming stimulus is not the same as the previous stimulus, because participants must freshly allocate attentional resources to the incoming stimulus. To compare previous and current events, participants must maintain the representation in working memory; this is more difficult for older adults.
Although our results for older adults may reflect an age-related change in maintenance in working memory, the present data are inadequate to support conclusions about characteristics of older adults with respect to the representation of stimuli in working memory. Therefore, further studies are needed to investigate this age-related change.
The appearance of P3b for frequently occurring standard stimuli, observed predominantly at the parietal site, did not reveal any differences between the social contexts for either age group, consistent with our previous findings [15]. In addition, the positive deflection was larger for older adults than for younger adults in the 400 -700 ms time window. These results suggest that the convergence of activities involved in stimulus evaluation was prolonged in the older adults.
The present study suggests that the presence of others did not affect the processing of targets, which were stimuli important to the participants themselves. As described above, for both age groups there were no differences between social contexts in P3b for targets. Thus older adults as well as young adults were able to allocate attentional resources to important events according to their own task representations.
Enhanced frontal positivity (no-go P3) for non-targets in joint-action settings was observed in the 200 -500 ms time window for older as well as younger adults. This is inconsistent with our prediction. The no-go P3, which reflects incorrect preparation, and then suppression, of motor responses, implies that older participants formed a representation of the partner's task, resulting in im-K. Kato, K. Yoshizaki Journal of Behavioral and Brain Science paired processing of the participant's own task-irrelevant non-targets. Interestingly, for older adults, a front-central dominant no-go P3 for non-targets explicitly appeared in the 400 -700 ms time window in the joint condition. This suggests that the inhibition of the partner's stimuli lasted longer for older adults than for younger adults.
The present results support previous findings [15] on the processing of other people's stimuli in young people. Kato et al. (2016) suggested that in situations in which a person is sharing tasks with others (i.e., the joint condition of the present study), two types of processing are being performed: matching incoming information with one's own task representation and matching it with the partner's task representation [15]. As can be seen in Figure 2, the non-target waveform of the younger adults is divided into first and second halves. Specifically, after the positive deflection returns toward baseline, the component once again shifts in a positive direction, a pattern mainly seen in the joint condition. According to Kato et al. (2016), the first and subsequent component correspond respectively to matching to the stimulus that the participant is in charge of and matching to the stimulus that the partner is in charge of [15]. Notably, the results for non-targets in the older adults were distinctly different from those in the younger adults. As can be seen in Figure 2, for older adults, one large positive deflection appeared from 200 ms to 700 ms in the non-target waveform.
Importantly, this deflection is regarded as a no-go P3, because it was dominant in the front-central area.
Note that the response decision can be made only by matching the incoming stimuli with one's own task representation in working memory. However, the present study demonstrated that the situation of task sharing (i.e., the joint condition) triggered matching to the partner's stimulus (i.e., non-targets) only for younger adults. The deterioration of attentional resources with aging likely led to the age-related difference in non-target processing. Participants should react to their own targets based on their own task representation. Since younger adults have sufficient attentional resources, they can afford to care about the other's action. Consequently, younger adults could allocate resources not only to their own target but also to their partner's target (their non-target). On the other hand, older adults do not execute the matching process to their partner's stimuli.
In spite of that, older adults seemed able to inhibit their partner's stimuli. This may reflect efficient use of limited resources by older people.
The disappearance of P3b for non-targets in joint settings may reflect the avoidance of errors by older adults. In other words, older adults do not allocate attention for further processing of irrelevant information to own task, irrespective of individual/social situations. Older adults who performed as well as the younger adults showed brain activity more extensive than that of younger adults, while older adults who performed more poorly produced brain activity at the same level as younger adults [33]. That is, processing stimuli in a mode similar to that of younger adults appears to result in a performance decline in older adults. Previous research has investigated whether older adults allocate more attentional resources to maintain their cognitive performance than do younger adults; this resource allocation peaks between the late 60s and early 70s for cognitively high performers [34]. Older adults in this age range are those who participated in the present research. The present findings of no age differences in P3b for targets may reflect this same attentional resource allocation that leads to the maintenance of task performance.
The limitation of this study is that we did not measure cognitive abilities such as working memory. There would be possibility that the age-related differences and non-differences obtain in the present study may depend on the individual differences of cognitive ability.

Conclusion
In summary, these ERP comparisons of younger and older adults performing a modified visual three-stimulus oddball task alone and together with another participant demonstrate characteristic usage of attentional resources with aging in social contexts. In agreement with our previous study, the younger adults could not only match incoming information with the other person's task representation but also inhibit incorrect preparation of motor responses instigated by representation of the other's task. Although older adults with cognitive decline could not match incoming information with the other's task representation, they could inhibit incorrect preparation of motor responses to non-targets in the social context. Taken together, although, other persons' task representation was formed irrespective of age in a joint setting, it is suggested that aging affects the way of processing based on the representation.