Neurophysiology of Human Perceptual Decision-Making.

The discovery of neural signals that reflect the dynamics of perceptual decision formation has had a considerable impact. Not only do such signals enable detailed investigations of the neural implementation of the decision-making process but they also can expose key elements of the brain's decision algorithms. For a long time, such signals were only accessible through direct animal brain recordings, and progress in human neuroscience was hampered by the limitations of noninvasive recording techniques. However, recent methodological advances are increasingly enabling the study of human brain signals that finely trace the dynamics of the unfolding decision process. In this review, we highlight how human neurophysiological data are now being leveraged to furnish new insights into the multiple processing levels involved in forming decisions, to inform the construction and evaluation of mathematical models that can explain intra- and interindividual differences, and to examine how key ancillary processes interact with core decision circuits.

Evidence Accumulation Rate Moderates the Relationship between Enriched Environment Exposure and Age-Related Response Speed Declines.

Older adults exposed to enriched environments (EEs) maintain relatively higher levels of cognitive function, even in the face of compromised markers of brain health. Response speed (RS) is often used as a simple proxy to measure the preservation of global cognitive function in older adults. However, it is unknown which specific selection, decision, and/or motor processes provide the most specific indices of neurocognitive health. Here, using a simple decision task with electroencephalography (EEG), we found that the efficiency with which an individual accumulates sensory evidence was a critical determinant of the extent to which RS was preserved in older adults (63% female, 37% male). Moreover, the mitigating influence of EE on age-related RS declines was most pronounced when evidence accumulation rates were shallowest. These results suggest that the phenomenon of cognitive reserve, whereby high EE individuals can better tolerate suboptimal brain health to facilitate the preservation of cognitive function, is not just applicable to neuroanatomical indicators of brain aging but can be observed in markers of neurophysiology. Our results suggest that EEG metrics of evidence accumulation may index neurocognitive vulnerability of the aging brain.Significance Statement Response speed in older adults is closely linked with trajectories of cognitive aging. Here, by recording brain activity while individuals perform a simple computer task, we identify a neural metric that is a critical determinant of response speed. Older adults exposed to greater cognitive and social stimulation throughout a lifetime could maintain faster responding, even when this neural metric was impaired. This work suggests EEG is a useful technique for interrogating how a lifetime of stimulation benefits brain health in aging.

Pharmacological Manipulations of Physiological Arousal and Sleep-Like Slow Waves Modulate Sustained Attention.

Sustained attention describes our ability to keep a constant focus on a given task. This ability is modulated by our physiological state of arousal. Although lapses of sustained attention have been linked with dysregulations of arousal, the underlying physiological mechanisms remain unclear. An emerging body of work proposes that the intrusion during wakefulness of sleep-like slow waves, a marker of the transition toward sleep, could mechanistically account for attentional lapses. This study aimed to expose, via pharmacological manipulations of the monoamine system, the relationship between the occurrence of sleep-like slow waves and the behavioral consequences of sustained attention failures. In a double-blind, randomized-control trial, 32 healthy human male participants received methylphenidate, atomoxetine, citalopram or placebo during four separate experimental sessions. During each session, electroencephalography (EEG) was used to measure neural activity while participants completed a visual task requiring sustained attention. Methylphenidate, which increases wake-promoting dopamine and noradrenaline across cortical and subcortical areas, improved behavioral performance whereas atomoxetine, which increases dopamine and noradrenaline predominantly over frontal cortices, led to more impulsive responses. Additionally, citalopram, which increases sleep-promoting serotonin, led to more missed trials. Based on EEG recording, citalopram was also associated with an increase in sleep-like slow waves. Importantly, compared with a classical marker of arousal such as α power, only slow waves differentially predicted both misses and faster responses in a region-specific fashion. These results suggest that a decrease in arousal can lead to local sleep intrusions during wakefulness which could be mechanistically linked to impulsivity and sluggishness.SIGNIFICANCE STATEMENT We investigated whether the modulation of attention and arousal could not only share the same neuromodulatory pathways but also rely on similar neuronal mechanisms; for example, the intrusion of sleep-like activity within wakefulness. To do so, we pharmacologically manipulated noradrenaline, dopamine, and serotonin in a four-arm, randomized, placebo-controlled trial and examined the consequences on behavioral and electroencephalography (EEG) indices of attention and arousal. We showed that sleep-like slow waves can predict opposite behavioral signatures: impulsivity and sluggishness. Slow waves may be a candidate mechanism for the occurrence of attentional lapses since the relationship between slow-wave occurrence and performance is region-specific and the consequences of these local sleep intrusions are in line with the cognitive functions carried by the underlying brain regions.

Contralateral delay activity is not a robust marker of cognitive function in older adults at risk of mild cognitive impairment.

Contralateral delay activity (CDA) has been proposed as a pre-clinical neural marker for mild cognitive impairment (MCI). However, existing evidence is limited to one study with a small sample size (n = 24). Our aim was to extend previous work by investigating the relationship between the CDA and MCI risk in a large sample of older adults (n = 76). We used a regression approach to determine whether (and when) CDA amplitude predicted MCI risk, as indexed by the Montreal Cognitive Assessment (MoCA). CDA amplitude from ~300-500 and ~800-900 ms predicted MoCA performance. However, significant effects were only observed for specific electrodes (P5/P6 and CP3/CP4, but not PO7/PO8) and the nature of the relationship between the CDA and MoCA scores differed across time and according to set size. Bayesian regression analysis indicated partial evidence in favour of the null hypothesis (BF10 values = 4-1.18). Contrary to previous results, our findings suggest that the CDA may not a robust marker of MCI risk. More broadly, our results highlight the difficulty in identifying at-risk individuals, particularly as MCI is a heterogeneous, unstable condition. Future research should prioritise longitudinal approaches in order to track the progression of the CDA and its association with cognitive decline in later life.

Antagonistic Interactions Between Microsaccades and Evidence Accumulation Processes During Decision Formation.

Despite their small size, microsaccades can impede stimulus detections if executed at inopportune times. Although it has been shown that microsaccades evoke both inhibitory and excitatory responses across different visual regions, their impact on the higher-level neural decision processes that bridge sensory responses to action selection has yet to be examined. Here, we show that when human observers monitor stimuli for subtle feature changes, the occurrence of microsaccades long after (up to 800 ms) change onset predicts slower reaction times and this is accounted for by momentary suppression of neural signals at each key stage of decision formation: visual evidence encoding, evidence accumulation, and motor preparation. Our data further reveal that, independent of the timing of the change events, the onset of neural decision formation coincides with a systematic inhibition of microsaccade production, persisting until the perceptual report is executed. Our combined behavioral and neural measures highlight antagonistic interactions between microsaccade occurrence and evidence accumulation during visual decision-making tasks.SIGNIFICANCE STATEMENT When fixating on a location in space, we frequently make tiny eye movements called microsaccades. In the present study, we show that these microsaccades impede our ability to make perceptual decisions about visual stimuli and this impediment specifically occurs via the disruption of several processing levels of the sensorimotor network: the encoding of visual evidence itself, the accumulation of visual evidence toward a response, and effector-selective motor preparation. Furthermore, we show that the production of microsaccades is inhibited during the perceptual decision, possibly as a counteractive measure to mitigate their negative effect on behavior in this context. The combined behavioral and neural measures used in this study provide strong and novel evidence for the interaction of fixational eye movements and the perceptual decision-making process.

Catecholamine Modulation of Evidence Accumulation during Perceptual Decision Formation: A Randomized Trial.

Recent behavioral modeling and pupillometry studies suggest that neuromodulatory arousal systems play a role in regulating decision formation but neurophysiological support for these observations is lacking. We employed a randomized, double-blinded, placebo-controlled, crossover design to probe the impact of pharmacological enhancement of catecholamine levels on perceptual decision-making. Catecholamine levels were manipulated using the clinically relevant drugs methylphenidate and atomoxetine, and their effects were compared with those of citalopram and placebo. Participants performed a classic EEG oddball paradigm that elicits the P3b, a centro-parietal potential that has been shown to trace evidence accumulation, under each of the four drug conditions. We found that methylphenidate and atomoxetine administration shortened RTs to the oddball targets. The neural basis of this behavioral effect was an earlier P3b peak latency, driven specifically by an increase in its buildup rate without any change in its time of onset or peak amplitude. This study provides neurophysiological evidence for the catecholaminergic enhancement of a discrete aspect of human decision-making, that is, evidence accumulation. Our results also support theoretical accounts suggesting that catecholamines may enhance cognition via increases in neural gain.

Visuospatial Asymmetries Arise from Differences in the Onset Time of Perceptual Evidence Accumulation.

Healthy subjects tend to exhibit a bias of visual attention whereby left hemifield stimuli are processed more quickly and accurately than stimuli appearing in the right hemifield. It has long been held that this phenomenon arises from the dominant role of the right cerebral hemisphere in regulating attention. However, methods that would enable more precise understanding of the mechanisms underpinning visuospatial bias have remained elusive. We sought to finely trace the temporal evolution of spatial biases by leveraging a novel bilateral dot motion detection paradigm. In combination with electroencephalography, this paradigm enables researchers to isolate discrete neural signals reflecting the key neural processes needed for making these detection decisions. These include signals for spatial attention, early target selection, evidence accumulation, and motor preparation. Using this method, we established that three key neural markers accounted for unique between-subject variation in visuospatial bias: hemispheric asymmetry in posterior α power measured before target onset, which is related to the distribution of preparatory attention across the visual field; asymmetry in the peak latency of the early N2c target-selection signal; and, finally, asymmetry in the onset time of the subsequent neural evidence-accumulation process with earlier onsets for left hemifield targets. Our development of a single paradigm to dissociate distinct processing components that track the temporal evolution of spatial biases not only advances our understanding of the neural mechanisms underpinning normal visuospatial attention bias, but may also in the future aid differential diagnoses in disorders of spatial attention.SIGNIFICANCE STATEMENT The significance of this research is twofold. First, it shows that individual differences in how humans direct their attention between left and right space reflects physiological differences in how early the brain starts to accumulate evidence for the existence of a visual target. Second, the novel methods developed here may have particular relevance to disorders of attention, such as unilateral spatial neglect. In the case of spatial neglect, pathological inattention to left space could have multiple underlying causes, including biased attention, impaired decision formation, or a motor deficit related to one side of space. Our development of a single paradigm to dissociate each of these components may aid in supporting more precise differential diagnosis in such heterogeneous disorders.

Abstract and Effector-Selective Decision Signals Exhibit Qualitatively Distinct Dynamics before Delayed Perceptual Reports.

UNLABELLED: Electrophysiological research has isolated neural signatures of decision formation in a variety of brain regions. Studies in rodents and monkeys have focused primarily on effector-selective signals that translate the emerging decision into a specific motor plan, but, more recently, research on the human brain has identified an abstract signature of evidence accumulation that does not appear to play any direct role in action preparation. The functional dissociations between these distinct signal types have only begun to be characterized, and their dynamics during decisions with deferred actions with or without foreknowledge of stimulus-effector mapping, a commonly studied task scenario in single-unit and functional imaging investigations, have not been established. Here we traced the dynamics of distinct abstract and effector-selective decision signals in the form of the broad-band centro-parietal positivity (CPP) and limb-selective ß-band (8-16 and 18-30 Hz) EEG activity, respectively, during delayed-reported motion direction decisions with and without foreknowledge of direction-response mapping. With foreknowledge, the CPP and ß-band signals exhibited a similar gradual build-up following evidence onset, but whereas choice-predictive ß-band activity persisted up until the delayed response, the CPP dropped toward baseline after peaking. Without foreknowledge, the CPP exhibited identical dynamics, whereas choice-selective ß-band activity was eliminated. These findings highlight qualitative functional distinctions between effector-selective and abstract decision signals and are of relevance to the assumptions founding functional neuroimaging investigations of decision-making. SIGNIFICANCE STATEMENT: Neural signatures of evidence accumulation have been isolated in numerous brain regions. Although animal neurophysiology has largely concentrated on effector-selective decision signals that translate the emerging decision into a specific motor plan, recent research on the human brain has isolated abstract neural signatures of decision formation that are independent of specific sensory and motor requirements. Here, we examine the functional distinctions between the two distinct classes of decision variable signal during decisions with deferred actions with and without foreknowledge of stimulus-effector mapping. We find salient distinctions in the dynamics of abstract versus effector-selective decision signals in the human brain, in terms of sustainment through response delays and contingency on foreknowledge of stimulus-response mapping.

Parsing the neural signatures of reduced error detection in older age.

Recent work has demonstrated that explicit error detection relies on a neural evidence accumulation process that can be traced in the human electroencephalogram (EEG). Here, we sought to establish the impact of natural aging on this process by recording EEG from young (18-35 years) and older adults (65-88 years) during the performance of a Go/No-Go paradigm in which participants were required to overtly signal their errors. Despite performing the task with equivalent accuracy, older adults reported substantially fewer errors, and the timing of their reports were both slower and more variable. These behavioral differences were linked to three key neurophysiological changes reflecting distinct parameters of the error detection decision process: a reduction in medial frontal delta/theta (2-7 Hz) activity, indicating diminished top-down input to the decision process; a slower rate of evidence accumulation as indexed by the rate of rise of a centro-parietal signal, known as the error positivity; and a higher motor execution threshold as indexed by lateralized beta-band (16-30 Hz) activity. Our data provide novel insight into how the natural aging process affects the neural underpinnings of error detection.

Transcranial direct current stimulation over right dorsolateral prefrontal cortex enhances error awareness in older age.

The ability to detect errors during cognitive performance is compromised in older age and in a range of clinical populations. This study was designed to assess the effects of transcranial direct current stimulation (tDCS) on error awareness in healthy older human adults. tDCS was applied over DLPFC while subjects performed a computerized test of error awareness. The influence of current polarity (anodal vs cathodal) and electrode location (left vs right hemisphere) was tested in a series of separate single-blind, Sham-controlled crossover trials, each including 24 healthy older adults (age 65-86 years). Anodal tDCS over right DLPFC was associated with a significant increase in the proportion of performance errors that were consciously detected, and this result was recapitulated in a separate replication experiment. No such improvements were observed when the homologous contralateral area was stimulated. The present study provides novel evidence for a causal role of right DLPFC regions in subserving error awareness and marks an important step toward developing tDCS as a tool for remediating the performance-monitoring deficits that afflict a broad range of populations.

The classic P300 encodes a build-to-threshold decision variable.

The P300 component of the human event-related potential has been the subject of intensive experimental investigation across a five-decade period, owing to its apparent relevance to a wide range of cognitive functions and its sensitivity to numerous brain disorders, yet its exact contribution to cognition remains unresolved. Here, we carry out key analyses of the P300 elicited by transient auditory and visual targets to examine its potential role as a 'decision variable' signal that accumulates evidence to a decision bound. Consistent with the latter, we find that the P300 reaches a stereotyped amplitude immediately prior to response execution and that its rate of rise scales with target detection difficulty and accounts for trial-to-trial variance in RT. Computational simulations of an accumulation-to-bound decision process faithfully captured P300 dynamics when its parameters were set by model fits to the RT distributions. Thus, where the dominant explanatory accounts have conceived of the P300 as a unitary neural event, our data reveal it to be a dynamically evolving neural signature of decision formation. These findings place the P300 at the heart of a mechanistically principled framework for understanding decision-making in both the typical and atypical human brain.

Characterising neural signatures of successful aging: Electrophysiological correlates of preserved episodic memory in older age.

While aging is associated with a gradual decline in memory, substantial preservation of function is observed in certain individuals and dissecting this heterogeneity is paramount to understanding successful aging. A cohort of elderly individuals were classified according to their level of memory preservation and administered a test of episodic memory in which participants were cued to learn or simply read each word and then to identify previously presented items in a delayed recognition phase. Mathematical modelling revealed that relatively preserved memory function was specifically linked to a faster rate of memorial evidence accumulation (drift rate). Analysis of event-related potentials at encoding revealed that high-performing elderly exhibited signals over parietal regions that discriminated between words to be learned vs. read for an additional 300-ms compared to young subjects suggesting a compensatory encoding mechanism that was absent in the low-performing group. At recognition, parietal signals associated with recollection processes discriminated previously learned words from read words in the young and high-performing old but not in low-performing old. These results reveal that successful aging is associated with specific adaptive neural markers during both encoding and retrieval.

Internal and external influences on the rate of sensory evidence accumulation in the human brain.

We frequently need to make timely decisions based on sensory evidence that is weak, ambiguous, or noisy resulting from conditions in the external environment (e.g., a cluttered visual scene) or within the brain itself (e.g., inattention, neural noise). Here we examine how externally and internally driven variations in the quality of sensory evidence affect the build-to-threshold dynamics of a supramodal "decision variable" signal and, hence, the timing and accuracy of decision reports in humans. Observers performed a continuous-monitoring version of the prototypical two-alternative dot-motion discrimination task, which is known to strongly benefit from sequential sampling and temporal accumulation of evidence. A centroparietal positive potential (CPP), which we previously established as a supramodal decision signal based on its invariance to motor or sensory parameters, exhibited two key identifying properties associated with the "decision variable" long described in sequential sampling models: (1) its buildup rate systematically scaled with sensory evidence strength across four levels of motion coherence, consistent with temporal integration; and (2) its amplitude reached a stereotyped level at the moment of perceptual report executions, consistent with a boundary-crossing stopping criterion. The buildup rate of the CPP also strongly predicted reaction time within coherence levels (i.e., independent of physical evidence strength), and this endogenous variation was linked with attentional fluctuations indexed by the level of parieto-occipital α-band activity preceding target onset. In tandem with the CPP, build-to-threshold dynamics were also observed in an effector-selective motor preparation signal; however, the buildup of this motor-specific process significantly lagged that of the supramodal process.

Pupil diameter covaries with BOLD activity in human locus coeruleus.

The locus coeruleus-noradrenergic (LC-NA) neuromodulatory system has been implicated in a broad array of cognitive processes, yet scope for investigating this system's function in humans is currently limited by an absence of reliable non-invasive measures of LC activity. Although pupil diameter has been employed as a proxy measure of LC activity in numerous studies, empirical evidence for a relationship between the two is lacking. In the present study, we sought to rigorously probe the relationship between pupil diameter and BOLD activity localized to the human LC. Simultaneous pupillometry and fMRI revealed a relationship between continuous pupil diameter and BOLD activity in a dorsal pontine cluster overlapping with the LC, as localized via neuromelanin-sensitive structural imaging and an LC atlas. This relationship was present both at rest and during performance of a two-stimulus oddball task, with and without spatial smoothing of the fMRI data, and survived retrospective image correction for physiological noise. Furthermore, the spatial extent of this pupil/LC relationship guided a volume-of-interest analysis in which we provide the first demonstration in humans of a fundamental characteristic of animal LC activity: phasic modulation by oddball stimulus relevance. Taken together, these findings highlight the potential for utilizing pupil diameter to achieve a more comprehensive understanding of the role of the LC-NA system in human cognition.

Prior probability cues bias sensory encoding with increasing task exposure.

When observers have prior knowledge about the likely outcome of their perceptual decisions, they exhibit robust behavioural biases in reaction time and choice accuracy. Computational modelling typically attributes these effects to strategic adjustments in the criterion amount of evidence required to commit to a choice alternative - usually implemented by a starting point shift - but recent work suggests that expectations may also fundamentally bias the encoding of the sensory evidence itself. Here, we recorded neural activity with EEG while participants performed a contrast discrimination task with valid, invalid, or neutral probabilistic cues across multiple testing sessions. We measured sensory evidence encoding via contrast-dependent steady-state visual-evoked potentials (SSVEP), while a read-out of criterion adjustments was provided by effector-selective mu-beta band activity over motor cortex. In keeping with prior modelling and neural recording studies, cues evoked substantial biases in motor preparation consistent with criterion adjustments, but we additionally found that the cues produced a significant modulation of the SSVEP during evidence presentation. While motor preparation adjustments were observed in the earliest trials, the sensory-level effects only emerged with extended task exposure. Our results suggest that, in addition to strategic adjustments to the decision process, probabilistic information can also induce subtle biases in the encoding of the evidence itself.

The speed of sight: Individual variation in critical flicker fusion thresholds.

The critical flicker fusion threshold is a psychophysical measure commonly used to quantify visual temporal resolution; the fastest rate at which a visual system can discriminate visual signals. Critical flicker fusion thresholds vary substantially among species, reflecting different ecological niches and demands. However, it is unclear how much variation exists in flicker fusion thresholds between healthy individuals of the same species, or how stable this attribute is over time within individuals. In this study, we assessed both inter- and intra-individual variation in critical flicker fusion thresholds in a cohort of healthy human participants within a specific age range, using two common psychophysical methods and three different measurements during each session. The resulting thresholds for each method were highly correlated. We found a between-participant maximum difference of roughly 30 Hz in flicker fusion thresholds and we estimated a 95% prediction interval of 21 Hz. We used random-effects models to compare between- and within-participant variance and found that approximately 80% of variance was due to between-individual differences, and about 10% of the variance originated from within-individual differences over three sessions. Within-individual thresholds did not differ significantly between the three sessions in males, but did in females (P<0.001 for two methods and P<0.05 for one method), indicating that critical flicker fusion thresholds may be more variable in females than in males.

Intracranial electroencephalography reveals effector-independent evidence accumulation dynamics in multiple human brain regions.

Neural representations of perceptual decision formation that are abstracted from specific motor requirements have previously been identified in humans using non-invasive electrophysiology; however, it is currently unclear where these originate in the brain. Here we capitalized on the high spatiotemporal precision of intracranial EEG to localize such abstract decision signals. Participants undergoing invasive electrophysiological monitoring for epilepsy were asked to judge the direction of random-dot stimuli and respond either with a speeded button press (N = 24), or vocally, after a randomized delay (N = 12). We found a widely distributed motor-independent network of regions where high-frequency activity exhibited key characteristics consistent with evidence accumulation, including a gradual buildup that was modulated by the strength of the sensory evidence, and an amplitude that predicted participants' choice accuracy and response time. Our findings offer a new view on the brain networks governing human decision-making.

Neurochemical enhancement of conscious error awareness.

How the brain monitors ongoing behavior for performance errors is a central question of cognitive neuroscience. Diminished awareness of performance errors limits the extent to which humans engage in corrective behavior and has been linked to loss of insight in a number of psychiatric syndromes (e.g., attention deficit hyperactivity disorder, drug addiction). These conditions share alterations in monoamine signaling that may influence the neural mechanisms underlying error processing, but our understanding of the neurochemical drivers of these processes is limited. We conducted a randomized, double-blind, placebo-controlled, cross-over design of the influence of methylphenidate, atomoxetine, and citalopram on error awareness in 27 healthy participants. The error awareness task, a go/no-go response inhibition paradigm, was administered to assess the influence of monoaminergic agents on performance errors during fMRI data acquisition. A single dose of methylphenidate, but not atomoxetine or citalopram, significantly improved the ability of healthy volunteers to consciously detect performance errors. Furthermore, this behavioral effect was associated with a strengthening of activation differences in the dorsal anterior cingulate cortex and inferior parietal lobe during the methylphenidate condition for errors made with versus without awareness. Our results have implications for the understanding of the neurochemical underpinnings of performance monitoring and for the pharmacological treatment of a range of disparate clinical conditions that are marked by poor awareness of errors.

Balancing true and false detection of intermittent sensory targets by adjusting the inputs to the evidence accumulation process.

Decisions about noisy stimuli are widely understood to be made by accumulating evidence up to a decision bound that can be adjusted according to task demands. However, relatively little is known about how such mechanisms operate in continuous monitoring contexts requiring intermittent target detection. Here, we examined neural decision processes underlying detection of 1 s coherence targets within continuous random dot motion, and how they are adjusted across contexts with weak, strong, or randomly mixed weak/strong targets. Our prediction was that decision bounds would be set lower when weak targets are more prevalent. Behavioural hit and false alarm rate patterns were consistent with this, and were well captured by a bound-adjustable leaky accumulator model. However, beta-band EEG signatures of motor preparation contradicted this, instead indicating lower bounds in the strong-target context. We thus tested two alternative models in which decision-bound dynamics were constrained directly by beta measurements, respectively, featuring leaky accumulation with adjustable leak, and non-leaky accumulation of evidence referenced to an adjustable sensory-level criterion. We found that the latter model best explained both behaviour and neural dynamics, highlighting novel means of decision policy regulation and the value of neurally informed modelling.

Confidence is predicted by pre- and post-choice decision signal dynamics.

It is well established that one's confidence in a choice can be influenced by new evidence encountered after commitment has been reached, but the processes through which post-choice evidence is sampled remain unclear. To investigate this, we traced the pre- and post-choice dynamics of electrophysiological signatures of evidence accumulation (Centro-parietal Positivity, CPP) and motor preparation (mu/beta band) to determine their sensitivity to participants' confidence in their perceptual discriminations. Pre-choice CPP amplitudes scaled with confidence both when confidence was reported simultaneously with choice, and when reported 1 second after the initial direction decision with no intervening evidence. When additional evidence was presented during the post-choice delay period, the CPP exhibited sustained activation after the initial choice, with a more prolonged build-up on trials with lower certainty in the alternative that was finally endorsed, irrespective of whether this entailed a change-of-mind from the initial choice or not. Further investigation established that this pattern was accompanied by later lateralisation of motor preparation signals toward the ultimately chosen response and slower confidence reports when participants indicated low certainty in this response. These observations are consistent with certainty-dependent stopping theories according to which post-choice evidence accumulation ceases when a criterion level of certainty in a choice alternative has been reached, but continues otherwise. Our findings have implications for current models of choice confidence, and predictions they may make about EEG signatures.