Brain components associated with EFs under depression

N100/N1, P200/P2, loudness dependence of auditory evoked potentials (LDAEP)

Different from P3, the N100 or N1 brain component, is not modality-independent. Thus, we can observe several differences in auditory N1 component versus the visual N1 component.

The visual N1 has diverse subcomponents. Earliest subcomponent peaks 100–150 ms after the appearance of stimulus and it is registered at anterior electrodes. Otherwise, the others N1 subcomponents peak later, around 150–200 ms post stimulus at posterior sites: parietal cortex and occipital cortex, and they are associated with attention and discriminative processing.

Concerning the auditory N1, it has also subcomponents. Among them, we can find a frontocentral component peaking around 75 ms, a central subcomponent peaking around 100 ms, and a lateral component peaking at 150 ms. These subcomponents seem to be related with attention too.

Regarding depression, studies frequently focus on auditory N1 component assessed through the oddball paradigm. The main reason to concentrate in auditory processing appears to be that primary auditory cortex receives many projections from neurons using a neurotransmitter closely linked to depression: serotonin [44, 45].

Neurotransmitters modulate ERPs, and ERPs’ studies show higher latency and lower amplitude in N1 under depressive conditions [46]. In addition, there are studies which identify a relationship between N1 and P3 components in depression. Increase in N1 latencies is associated with a decrease in P3 amplitude, what could affect discrimination and information processing. And alteration in attention and perception is likely to affect later cognitive processes related to updating of new information in working memory [46–48].

Besides, a specific feature of auditory processing is being studied in the last years. It refers to the relationship between the N1 and P2 components. P2 component is associated with pre-attentive auditory responses. The change in amplitude of the auditory N1/P2 component (difference between N1 and P2) in response to diverse stimulus intensities, also named LDAEP (loudness dependence of auditory evoked potentials), seems to be and indirect indicator of central nervous system serotonin function [44].

Nonetheless, results were inconsistent concerning LDAEP in depression. Studies comparing LDAEP between depressed persons and non-depressed controls show an increase LDAEP in depressed in some cases, a decrease in other cases, or no differences between groups. Reasons could be related with heterogeneity of depression, experimental designs, or the complexity of the physiology of depression. Although serotoninergic dysfunction in the central nervous system is considered to be one of the major organic pathophysiological factors of depression, further neurotransmitters are implicated in depression, such as dopamine, glutamate or noradrenaline among others.

Additionally, we can find a variety of procedures for measuring LDAEP that have been used in the experiments. Often, LDAEP is defined as an amplitude change in the auditory N1/P2 component in response to a diversity of stimulus intensities, as we have previously mentioned. But LDAEP could also be measured as amplitude change in N1, P1, P2 and P1/N1 component in response to stimulus intensities [31].

From a cognitive-behavioural perspective, N1 and P2 are not only associated with pre-attentive and discriminative functions, but also with EFs. Poor performance of No-go in Go/No-go paradigms (related to EFs, specifically with inhibitory control), could be associated with low serotonin function. As mentioned before, LDAEP could work as an indirect indicator of serotonin levels, higher LDAEP is linked to low serotonin. In fact, impulsive persons show stronger LDAEP, difficulties controlling the inhibitory response for No-go trials, and higher emotional sensitivity associated with depressive states [49, 50].

But further research is needed to clarify the variety of results in LDAEP in relation with depression, and whether inhibitory control is directly related with N1/P2 or it is the result of the effect of attentional deficits on other components.

Mismatch negativity (MMN)

Continuing with the auditory response, it is relevant to mention the MMN. This brain component consists in a negative wave that use to peak between 150–250 ms, at central and fronto-central locus, after the presence of an occasional, deviant, or mismatching stimulus, appearing in a succession of identical stimuli (e.g., a sequence of 800-Hz tones and occasionally 1,300-Hz tones).

The MMN is considered an automatic brain response elicited even when the person is not involved actively in the auditory task, but listening the sounds while performing another activity, such as reading. In addition, MMN can also be registered under other sensory modalities (i.e., somatosensory, olfactory, visual modality or even in audio-visual tasks) [51–53].

Although some studies have claimed that the MMN is not different from the N1 wave, and they consider the MMN response a product of a posterior subcomponent of N1 peaking around 85 ms (“N1p”) and an anterior subcomponent of N1, peaking about 150 ms after stimulus (“N1a”) [54], extensive research seems to demonstrate that MMN must be considered a differentiate component from N1 due to several facts.

First, the MMN latency and duration do not match the latency and duration of N1a; secondly, the MMN response could be elicited by deviants that do not produce N1 response; third, when an experiment presents a repeated sequence of tones increasing or decreasing in frequency (standards), a different MMN response could be registered, even if the settings do not allow the adaptation of N1-generating neurons that appears with deviants occurring in a sequence of similar-frequency repetitive stimuli.

Then, the MMN scalp distribution and generator locus are different from those of N1a. Auditory-cortical N1 subcomponents show larger amplitude on the hemisphere contra-lateral to the stimulated ear, while the MMN changes due to frequency, duration, and intensity of stimuli, are right-hemispheric predominant. Moreover, the MMN has a second main generator in the frontal lobes, with higher amplitudes at these sites, that cannot be observed in N1a topography.

Lately, an important difference between MMN and N1 components, concerns the effects on N1a after experimental changes, which do not mimic the effects on the MMN response. It has been explored analysing diverse effects of pharmacological interventions on N1 and MMN, or how certain lesions may eliminate the MMN but do not affect the N1 activity [55].

Regarding the MMN functions, this automatic brain response is considered a pre-attentive index of deviance detection occurring not only when the person is concentrated in the task, but also when subjects perform another activity that is not related with the auditory stimulus or another passive task [56].

It has been stated that brain may elicit MMN because it can recognize a distinct stimulation. Thus, information has been previously stored in sensory memory, and a deviant stimulus is being compared with the other sounds. Then, the brain, cognitive, and perceptive systems, can detect the discrepancy between the neural representation of the regular stimuli and the representation of the given mismatch stimulus, producing the MMN [57]. Thus, MMN could be considered an index of change detection occurring unintentionally and automatically, and also, an index associated with auditory sensory memory.

However, the auditory memory cannot be understood as a unique process, but rather as a domain including several processes. In regard to MMN, at least four steps can be identified: (1) analysis, categorization and classification of incoming input, (2) integration of information into a logical model, (3) predictions about future stimulations based on the built model, and (4) comparison of representation, between previous organized and stored information, with new information [58].

Additionally, some authors assert that the MMN could be described as a stage of the distraction effect or the involuntary capture of attention, where the sound that violates expectations trigger initially a MMN, then, in a second stage, P3 reflects the orientation of attention toward the unexpected stimulus, and finally, the reorienting negativity (RON) response appears, a negative wave occurring 500 ms after stimulus, linked to the turning attention back to the primary task [59, 60].

Nonetheless, some studies have reported contradictory results, showing that N1/MMN, P3a and RON do not form a coupled chain reflecting the three stages of auditory distraction. Indeed, P3 could appear without eliciting MMN and consequently RON response [61].

About EFs, MMN appears to be not only related with an automatic pre-attentive response or an involuntary capture of attention by deviant auditory stimuli, but also with a voluntary effort related to attention switching or mental shift.

Previous experiments have reported that the MMN originates from auditory cortex on the supratemporal region and have at least two sources. One in the right lateral temporal cortex and another in the right frontal cortex [62]. The prefrontal cortex has a relevant role in controlling the direction of attention, so it could represent the initiation of the involuntary orienting of attention to a previous detected changed by the auditory-cortex MMN. Thus, the prefrontal activation of MMN, occurring later than the temporal source of MMN, could be understood as a process of attention switching [63].

However, Toyomaki et al. [64] found a correlation between low MMN amplitude and poor EFs performance in schizophrenia patients, using non-auditory tasks, specifically using the WCST, the Trail Making Test, and the Stroop Test to assess cognitive control, cognitive witching, and response inhibition. And they concluded that the frontal component of the MMN may be associated with conscious perception of attention switch to stimulus change.

Future research needs to clarify whether the frontal source of MMN is related with a voluntary or an involuntary process, associated exclusively with attention or also with executive domains and other networks.

Respecting depression, a recent meta-analysis including 13 studies out of 438, based on auditory paradigms, has shown that amplitude and latency of MMN regarding duration but not frequency deviants, were significantly impaired under depression in comparison with healthy controls, registering lower MMN amplitudes and prolonged MMN latencies [65].

Further projects have tried to analyse how a variety of deviant types might elicit a different MMN waves under depression, registering an increase on MMN amplitudes following tones that deviated in intensity and location [66].

Conclusively, although many studies report a decrease of MMN amplitudes and longer latencies under depression, results are inconsistent due to the variety of stimuli features, methodological issues and the severity of depression [31], and more research is required to evaluate the relationship among depression, MMN and EFs.

Feedback processing, error processing

Lately, it is important to highlight the brain activity associated with feedback processing and error processing. Once the person has provided a response, the evaluation of the feedback (correct or incorrect answer), and the ability to update the information, conflict monitoring, error processing, and the decision about continuing a response criterion or change it, are essential EFs that allow persons to adapt constantly to the environmental demands, and perform efficiently daily life activities.

ERN or error negativity (EN) consists in a negative deflection peaking around 100 ms of error commission at frontal and central electrodes, during simple decision tasks [67, 68].

Although an association between depression and ERN has been demonstrated [69, 70], it is not clear the direction of it. Some experiments have obtained enhanced ERN amplitudes in people with depression, while other show smaller ERN amplitudes under depression [71, 72].

Traditionally, depression has been characterized by excessive sensitivity to negative information. There is a tendency to amplify mistakes and consequences, a bias towards negative experiences, and a higher sensitivity to NF.

Authors who find an enhanced amplitude of the ERN component, explain it due to the hypersensitivity to errors that depressed persons show in early stages of error detection and error processing. Depression triggers an exaggerated early error response that require higher neural, cognitive and perceptive resources, generating subsequently difficulties to adapt to new situations or stimuli.

This idea is supported by neuro-structural and functional data which recognize that the anterior cingulate cortex (ACC) is likely the generator of ERN. This brain structure shows significant increase of metabolic activity under depression, coupled with a decreased activity in prefrontal regions, what could be related with the alterations observed in EFs [71, 73, 74].

However, some studies do not replicate these results. They find smaller (less negative) ERN amplitudes using different types of tasks, accounting for errors of choice (pressing an incorrect response button in an Erikson Task), and errors of commission (pressing a button when one is not supposed to, in a Go/No-go Task), that could be linked to difficulties in the evaluation of NF under depression.

In these cases, the results are in line with deficits in prefrontal functions and executive components of goal-directed behaviour, strategic reasoning, and cognitive control under depression. But functional and neuroimaging studies show decrease activation not just at prefrontal cortex but also at ACC, what contradicts previous results [75, 76].

Thus, the data regarding ERN are contradictory, and it is necessary to perform replication studies to understand the differential amplitudes of this brain component, and the neural networks implicated in error processing under depression. In addition, it is required to clarify whether depression severity and heterogeneity, together with comorbidity, could explain partly the differential results.

Moreover, wrong responses or errors are associated not only with ERN but also with a positive deflection peaking around 200 ms after the erroneous answer at posterior midline scalp distribution. This positive wave is named error positivity (Pe). Pe is independent of ERN and is linked with a later aspect of error processing. It has been related with conscious error awareness, orienting response to errors, cognitive control, and an emotional assessment process or motivational significance of errors, where mistakes are evaluated depending on the significance for the individual [67, 77, 78].

Further investigation needs to explain the relationship between Pe and depression, but several studies have found smaller Pe amplitudes under depression, what could be explained because depression could lead to a difficulty to be aware of errors and differentiate wrong and correct responses. A second explanation related to the emotional role of Pe is that a lower Pe in depressive persons is due to indifference to commit mistakes [72], what could be connected with anhedonia and apathy, indirectly associated with EFs [79, 80].

However, it is necessary to clarify why some experiments show a decrease in Pe under depression, coupled in some cases with an enhancement of ERN, and in other situations, with a reduction of ERN, and to understand its functional and clinical significance.

Pe has also been coupled with P300. Findings show reduced Pe and covariation of P3 amplitude under depression, but they seem to reflect distinct responses or manifestations of the same impairment in depression: reduced awareness and attentional allocation to stimuli [81, 82].

Similar to ERN, it is possible to identify a negative wave related to correct responses, the correct response negativity (CRN), peaking around 50 ms after the positive feedback at fronto-central regions, indicating a basic response monitoring process.

Regarding depression, results about CRN are also mixed, showing studies with a CRN amplitude reduction, experiments where CRN amplitudes are more negative in depression, or cases with similar CRN activity in experimental and control groups.

Besides, some studies report increased CRN under severe depression, what could be explained by an increased error monitoring on correct choices due to a tendency of depressed individuals to be uncertain about the veracity of their correct answers or the expectancy for committing more errors [83].

Nonetheless, these results need to be more investigated, because some analysis demonstrated that CRN could be considered an ERN, and the ERN may not be just understood in terms of error detection but from a broad perspective [84].

Continuing with the feedback processing, and regarding the WCST, this task requires subjects to stablish, maintain or shift sorting rules depending on the received feedback. These abilities are related with the capacity of individuals for adapting to changing environments, what is named set-shifting, mind-switching or cognitive flexibility, key components of EFs.

According to previous research, we can distinguish two moments of feedback in the WCST. The first one refers to NF coming after a sequence of positive feedback, what means that the subject must change the classification criteria. The second moment of feedback comes after the switching rule has been applied. Here, the participant could receive second-learning NF, so the new criteria is not correct, and it is required a second change; or the person can get positive feedback, indicating that the new sorting rule is right.

It is recognized that feedback has two components. The feedback could indicate whether the answer is correct or incorrect. This component is the valence of the feedback. Additionally, we can learn something else from the feedback beyond the valence, we can infer information about the rules and its implications. This component is named the informative value of feedback.

Thus, in the classic WCST we differentiate two moments: the first NF and possibly a second-learning NF; and two feedback components: the valence (correct or wrong response), and the informative value (information and implications of the rules and criteria).

Once a person gets feedback with a negative valence (incorrect answer), the person stays in the first moment or stage of NF. The person knows that it is not possible to continue with the same sorting criteria, and it should be changed according to the rules in order to get positive feedback. All these processes are related with the informative value of the NF, and this first moment could be identified as a rule-switch stage.

Then, the person tries another criterion and gets feedback (positive or negative) that is going to help to learn the new classification rule. This second moment is also called rule-learning stage. In both stages is possible to obtain NF and they are frequently denominated Switch-NF and Learn-NF.

Concerning neural activity, several studies have found diverse brain activity associated with both moments and components. Data shows that P3 is linked to the information value of feedback, and it increases in amplitude when it is necessary to extract more information from the feedback, what could be related with information updating [85, 86].

However, no information was found about the ERN and CRN for the different valence and moments of feedback in general, and specifically for the NF. And no information was found about Switch-NF and Learn-NF under depression either.

Finally, some studies using WCST, have analysed the brain components associated with the two types of errors that could be made in this task: perseverative errors and non-perseverative errors. Perseverative errors refer to the perseverance or persistence to respond according to an incorrect criterion or stimulus characteristics.

On the other hand, non-perseverative errors refer: firstly to errors that happen immediately after the sorting rule changes and indicate that it is necessary to adjust the response to a new matching criterion, and secondly, non-perseverative errors refer to failures to maintain mind-setting responding to the same sorting rule, due to the inability to maintain the focus or to inhibit the distracting interference of the other stimuli presented.

Depression characterizes by executive deficits assessed with the WCST, in terms of perseverative errors and non-perseverative errors, what could lead to poorer clinical outputs [87, 88].

If a person tends to commit perseverative errors, it likely means that experience difficulties during both moments of NF: the first NF and the second learning NF (Switch-NF and Learn-NF), and difficulties to understand the informative value of the NF.

The neural bases of this phenomenon need to be further explored, but some experiments have analysed the underlying mechanisms of perseverative and non-perseverative errors under healthy conditions, and they found distinct patterns of brain activation evoked by both types of errors.

Perseverative errors evoked larger P3b than non-perseverative errors or distractions, mainly at Pz location, what could suggest different brain processing in later stages at posterior regions as consequence of different early frontal activity. Moreover, at fronto-central regions, P2 showed higher amplitudes on distractions than on perseverative errors, and the parieto-occipital N1 was absent at perseverative errors.

It may reflect that perseverative errors are the result of alteration of fronto-extriatal network, because the absence of extrastriate N1 precedes the reduction of P2 at frontal areas, indicating troubles in visual attention to stimuli; whereas, distractions or non-perseverative errors seems to be more related with an inability to inhibit interfering information [42]. Nonetheless, further research should be conducted, specially, to understand the error and feedback processing under depression.