Electrophysiological markers of cognitive control in the Stroop task: analysis of event-related potentials

Shlyakhov Ivan Sergeevich ORCID: 0000-0001-7801-382X Assistant; Department of Psychology; V.A. Almazov National Medical Research Center Postgraduate student; Faculty of Psychology; Saint Petersburg State University 199225, Russia, St. Petersburg, Vasileostrovsky district, Vadim Shefner str., 14, room 1, building 1 i.shlyakhov@spbu.ru Gorbunov Ivan Anatol'evich ORCID: 0000-0002-7558-750X PhD in Psychology Senior Lecturer; Faculty of Psychology; St. Petersburg State University 199178, Russia, St. Petersburg, Vasileostrovsky district, line 8-ya.O., 77 i.a.gorbunov@spbu.ru

Dakhnovskaya Mariya Sergeevna ORCID: 0009-0007-3698-3394 Student; Faculty of Psychology; Saint Petersburg State University 199155, Russia, St. Petersburg, Vasileostrovsky district, Zheleznovodskaya str., 15 letter B st118249@student.spbu.ru

Introduction. Cognitive control is a set of cognitive processes that provide flexibility of thinking, regulation of behavior and adaptation to changing environmental conditions ^[5]. These processes include searching for conflicting information, suppressing interference, and resolving conflicts that require attention and conscious control.

One of the most studied paradigms for assessing cognitive control is the Stroop task ^[1], in which subjects need to name the color of a word, ignoring its semantic meaning. The conflict between the automated process of reading a word and the controlled process of determining color requires significant involvement of executive functions. Studies using neuroimaging and electroencephalography (EEG) techniques have shown that performing the Stroop task is associated with the activity of the anterior cingulate cortex (ACC), dorsolateral prefrontal cortex (dlPFC) and other areas responsible for cognitive control ^[4].

Recordings of evoked potentials (VP) make it possible to study the temporal dynamics of cognitive control in the Stroop problem with high accuracy. In particular, three key components of VP were identified related to various aspects of cognitive control: frontocentral N2, reflecting conflict monitoring processes ^[4], N400, associated with interference suppression ^[10], and a late positive component (LPC), which is associated with conflict resolution processes ^[7].

Thus, the study of the Stroop problem using EEG provides important data on the temporal and spatial organization of cognitive control, allowing us to clarify the mechanisms underlying the processing of conflicting information.

Electroencephalography (EEG) is one of the key methods of studying cognitive control, which allows recording the electrical activity of the brain with high time resolution. In particular, VP analysis makes it possible to identify specific neurophysiological markers reflecting the processes of conflict monitoring, interference suppression, and conflict resolution.

One of the most reliable VP markers of cognitive control is the N2 component, which manifests itself in the range of 200-350 ms after presentation of the stimulus. The main neural generators of N2 are considered to be the anterior cingulate cortex (ACC) and the dorsolateral prefrontal cortex (dlPFC) ^[12].

Another important VP component is N400, which plays a key role in suppressing interference in cognitive control tasks, including the Stroop task. In tasks requiring interference suppression, N400 is most pronounced in the central-parietal regions and is associated with the activity of the anterior cingulate and prefrontal cortex ^[9].

The late positive component (LPC) occurs in the range of 500-800 ms after the presentation of the stimulus and is associated with conflict resolution processes ^[10]. LPC most often manifests itself as a positive amplitude shift in the parietal and central areas of the scalp, which indicates the involvement of conscious mechanisms of cognitive control and information processing after conflict ^[8]. In Stroop's task, LPC enhancement is observed in response to conflicting stimuli and reflects the mobilization of cognitive resources to choose the correct response ^[15].

The study of cognitive control in the Stroop problem using electroencephalography (EEG) is an important methodological approach to the study of neurophysiological mechanisms of conflict information processing. Despite a significant amount of work devoted to this topic, there are still a number of unresolved issues regarding the exact temporal and spatial characteristics of the brain structures involved in this process, as well as the relationship between the various components of evoked potentials (VP).

The purpose of this study is to study the features of evoked potentials in the Stroop task with an emphasis on three key components of the VP: N2, related to conflict monitoring, N400, reflecting interference suppression, and the late positive component (LPC) associated with conflict resolution.

Based on previous research, the following hypotheses are put forward in this paper:

1. The N2 component will manifest itself with a greater amplitude in conflict conditions, which indicates activation of the anterior cingulate cortex (ACC) when interference is detected ^[20].

2. The N400 component will exhibit a more pronounced negative amplitude in response to conflicting stimuli, which is associated with the activity of the prefrontal cortex and anterior cingulate cortex in suppressing automated responses ^[17].

3. The late positive component (LPC) will be more pronounced in terms of conflict resolution, which indicates the mobilization of cognitive resources to perform a correct response ^{[10; 8]}.

This study will clarify the mechanisms of cognitive control in the Stroop task, as well as complement existing models describing the temporal dynamics of conflict monitoring, interference suppression, and conflict resolution ^[3].

Materials and methods. The study collected data from 36 people aged 18 to 54 (average age 36 years), 17 of them women and 19 men. The subjects had no diagnosed neurological or mental illnesses. Written informed consent to participate in the study was also obtained from each subject.

Based on the analysis of data obtained from 36 participants, the normality of the distribution was checked using the Shapiro-Wilk test and the significance of the differences was assessed using the Kraskel-Wallis test.

Results. In this study, VP was analyzed when performing the Stroop task to assess the neurophysiological mechanisms of cognitive control. The main focus was on components N2, N400, and the late positive component (LPC), as they reflect the processes of conflict monitoring, interference suppression, and conflict resolution, respectively.

Verification of the normality of the data distribution using the Shapiro-Wilk test showed a significant deviation from the normal distribution in all analyzed channels (p < 0.05). This confirmed the need for nonparametric analysis methods. The results of the Kruskal-Wallis test revealed statistically significant differences (p < 0.05) in the amplitudes of the VP components between the conditions (correct and incorrect answers), which indicates differences in the processing of conflicting information depending on the success of the task.

Table 1

Significance assessment (Kruskal-Wallis test)

Fig. 1. Topograms when guessing a stimulus

When guessing the stimulus, the evoked potentials demonstrated characteristic features associated with successful cognitive processing of information. The P300 component, manifested in the range of 300-500 ms, was most pronounced in the central, parietotemporal, and parietal regions. For example, a positive wave with an amplitude of 1.31–1.80 MV was observed in the T5 channel in the range of 472-520 ms. This indicates the successful completion of a cognitive task.

Fig. 2. Evoked potential for the T5 channel when guessing a stimulus

The late positive component (LPC) associated with conscious information processing was also evident when guessing a stimulus, but with a lower amplitude compared to incorrect responses. This may indicate a lower cognitive load when processing relevant information.

When the stimulus was not anticipated, the evoked potentials showed more pronounced changes associated with the detection of errors and the processing of conflicting information. At the early stages of processing, the ERN (Error-Related Negativity) component was observed in the range of 50-150 ms, which was most pronounced in the central regions (for example, in the Cz channel, the amplitude was 1.62–0.32 Mv).

Fig. 3. Topograms when a stimulus is not predicted

The N200 component associated with the processes responsible for enhancing cognitive control was manifested in the range of 200-300 ms and was especially pronounced in the temporal regions (for example, in the T5 channel the amplitude was -6 MV).

Fig. 4. Evoked potential for the T5 channel when the stimulus is not anticipated

Late components, such as LPC, manifested with a higher amplitude when the stimulus was not anticipated (for example, in the Cz channel the amplitude was 2.04–2.15 Mv in the range of 520-712 ms). This may be due to an increase in cognitive load for conscious control of errors that occur during the task.

Frontal waves (N400) associated with semantic conflict processing were also more pronounced when the stimulus was not anticipated. For example, a negative wave with an amplitude of -3.83-3.09 MV was observed in the F3 channel in the range of 328-376 ms, which indicates the difficulties caused by interference in processing conflicting information.

Fig. 5. Evoked potential for the F3 channel when the stimulus is not anticipated

A comparison of evoked potentials in guessing and non-guessing stimuli revealed significant differences in the amplitude and latency of key components. When guessing a stimulus, the positive components are more pronounced, reflecting the successful completion of a cognitive task and the effective distribution of attention. At the same time, when a stimulus is not predicted, more pronounced negative components are observed, such as ERN, N200 and N400, which is associated with the process of detecting errors in processing conflicting information and with the activation of cognitive control.

Fig. 7. Comparison of evoked potentials for the Cz channel when guessing and not guessing a stimulus

The late positive component (LPC) was more pronounced when the stimulus was not anticipated, indicating the need for information processing and error awareness. This confirms that not anticipating a stimulus requires greater involvement of cognitive resources.

The topographical differences also highlight the difference between guessing and not guessing a stimulus. When guessing the stimulus, activation was more pronounced in the central and parietal regions, whereas when not guessing the stimulus, the most significant changes were observed in the frontal and frontal regions, which is associated with the processing of conflicting stimuli.

Thus, differences in evoked potentials confirm that guessing and not guessing a stimulus activate various neural mechanisms related to cognitive control, attention, and information processing.

Discussion. The results of this study confirm the key role of cognitive control in conflict information processing in the Stroop task. The VP analysis identified three main components — N2, N400, and LPC — each of which is associated with certain aspects of cognitive control, including conflict monitoring, interference suppression, and conflict resolution. The data obtained are consistent with the results of previous studies, indicating the involvement of the anterior cingulate cortex and the dorsolateral prefrontal cortex (dlPFC) in the regulation of these processes ^[13].

Component N2 demonstrated the expected amplification of the amplitude in conditions of high conflict, which confirms its connection with conflict monitoring and the involvement of the control panel in the detection of conflicting information ^[19]. This result is consistent with the cognitive control model, according to which the ACC plays a leading role in error detection and attention management in conditions of cognitive conflict ^[6].

The N400 component, which manifests itself as an amplified negative wave in the central-parietal region, has confirmed its role in suppressing interference between verbal and color information. This effect can be interpreted as an indicator of the suppression of automated reactions requiring the processing of conflicting stimuli. This is consistent with the data from neuroimaging studies demonstrating the activity of the anterior cingulate and prefrontal cortex when performing tasks requiring interference control ^[18].

LPC showed a more pronounced amplitude in the conditions of successful conflict resolution, which indicates the mobilization of cognitive resources to choose the correct response. This result is consistent with the hypothesis that LPC reflects the conscious processing of information and the involvement of executive functions in decision-making ^[16]. Similar conclusions have been drawn in studies using combined EEG-fMRI methods confirming the role of LPC in conflict processing and behavior adaptation ^[13].

The data obtained are also consistent with the theories of two-phase cognitive control, according to which reactive control (associated with N2) precedes proactive control associated with interference suppression (N400) and final conflict resolution (LPC) ^[5]. In particular, our results confirm that effective performance of the Stroop task requires the interaction of several levels of cognitive control, each of which involves specific neural networks.

The results of the study confirm the theory of two-phase cognitive control, according to which the early stage (associated with N2) is responsible for conflict detection, the middle stage (N400) is associated with interference suppression, and the late stage (LPC) reflects conflict resolution and adaptation processes ^[14]. These conclusions are consistent with functional neuroimaging data demonstrating the activity of the corresponding cortical regions when performing tasks with a high level of interference ^[2].

Future research may focus on individual differences in cognitive control, as well as on the influence of factors such as bilingualism or age on the effectiveness of interference suppression. In addition, a promising direction is the analysis of the temporal dynamics of VP components in the context of various cognitive control strategies ^[11].