Skip to main content
NCBI home page
NIHPA Author Manuscripts logoLink to NIHPA Author Manuscripts
. Author manuscript; available in PMC: 2024 Jun 21.
Published in final edited form as: Neurosci Lett. 2023 May 10;808:137294. doi: 10.1016/j.neulet.2023.137294

Late ERP correlates of confidence for auditory categorization of complex sounds

Alexandria C Zakrzewski a,*, Brian Maniscalco b, Matthew G Wisniewski a
PMCID: PMC10330643  NIHMSID: NIHMS1900857  PMID: 37172774

Abstract

Recent research suggests that confidence judgments relate to the quality of early sensory representations and later modality independent processing stages. It is not known whether the nature of this finding might vary based on task and/or stimulus characteristics (e.g., detection vs. categorization). The present study investigated the neural correlates of confidence using electroencephalography (EEG) in an auditory categorization task. This allowed us to examine whether the early event-related potentials (ERPs) related to confidence in detection also apply to a more complex auditory task. Participants listened to frequency-modulated (FM) tonal stimuli going up or down in pitch. The rate of FM tones ranged from slow to fast, making the stimuli harder or easier to categorize. Tone-locked late posterior positivity (LPP) but not N1 or P2 amplitudes were larger for (correct-only) trials rated with high than low confidence. These results replicated for trials presenting stimuli at individually identified threshold levels (rate of change producing ~71.7% correct performance). This finding suggests that, in this task, neural correlates of confidence do not vary based on difficulty level. We suggest that the LPP is a task general indication of the confidence for an upcoming judgment in a variety of paradigms.

Keywords: Confidence, metacognition, ERP, EEG, late posterior positivity, P3

1. Introduction

Metacognitive judgments such as confidence can guide our decisions about the sensory information in the world around us. For example, confidence in the detection and identification of an ambulance siren can impact driving decisions. If one is not confident that an ambulance siren is present, one may listen longer to collect more information before pulling over to the roadside. Otherwise, false alarms to non-siren sounds (e.g., birds, radio, etc.) might lead to excessive stopping.

Several studies have shown that confidence judgments track the accuracy of perception [1], of memory (e.g., [2]), and of both [3–8]. Relatedly, event-related potential (ERP) work has shown that late ERP components correlate with decision confidence. In particular, late posterior positivities (LPPs) reminiscent of the P3b are larger in amplitude on high- compared to low-confidence trials (e.g., [9–11]). However, most of this work has been done in the visual domain with categorization or memory tasks. Only a few have assessed whether or not neural correlates of metacognitive judgments generalize across sensory modalities and task complexity (e.g., [12,13]). Squires et al. [14] and, recently, Zakrzewski et al. [15] have shown the amplitude of relatively early auditory ERP components (N1 and P2) track confidence ratings along with the LPP responses typically observed in visual studies. This could be taken to suggest that the stimulus processing associated with metacognitive confidence ratings can be based on the quality of early sensory representations and later modality independent processing stages. Currently, it is not known whether the nature of this finding is based in sensory modality (auditory vs. visual) or task complexity (detection vs. categorization/memory). Indeed, a major challenge in studying the neuroscience of metacognition has been that metacognitive processes often vary with changes in task and stimulus characteristics (for discussion, see [16,17]). The difference between the auditory detection work and the visual ERP work in the literature is a prime example of this.

The present study explored the ERP correlates of confidence judgments in an auditory categorization task. On each trial in the experiment, listeners were presented with a single frequency modulated tonal stimulus. This modulation could be from low- to high-frequency (i.e., going up in pitch over time), or from high- to low-frequency (i.e., going down in pitch). The rate of frequency modulation (FM) ranged from relatively fast (easy trials) to slow (difficult trials). Participants made confidence ratings after each “up” vs. “down” decision. The novelty of the categorization task allowed us to examine whether the early ERP correlates of confidence judgments in detection also apply to a more complex auditory task. The FM rate manipulation allowed us to assess whether ERP relationships existed within several levels of difficulty (correlate exists even within same stimulus level) or were primarily driven by task difficulty itself (ERP component amplitude parallels changes in FM rate). The study therefore stands to inform how task complexity (through comparisons to previous research) and difficulty impact the processes on which metacognitive judgments are based.

2. Pilot Experiment

A behavioral pilot experiment was conducted to identify octaves/s thresholds for making the up vs. down FM sweep categorization. This data was used to select a range of FM sweep rates for the main experiment.

2.1. Material and Methods.

2.1.1. Participants.

Twenty-seven individuals (21 females and 6 males; ages 18 – 26. M = 19.33 SD = 2.32) participated for course credit from Kansas State University (KSU). All participants in the pilot experiment reported normal hearing. All 27 participants of the pilot experiment were included in the final analysis.

2.1.2. Stimuli.

Single frequency-modulated (FM) tonal sweeps were used as stimuli (cf. [18]). FM sweeps share many of the features important for categorizing real-world relevant sounds but are not generally related to pre-existing categories and biases that would be present for speech and/or environmental sounds. All FM sweeps were 100 ms in duration and swept up or down in frequency with the lowest frequency at 1500 Hz (for octaves/s, see section 2.1.4).

2.1.3. Equipment.

Participants sat in a sound attenuating booth while hearing sounds over Sennheiser HD-280 closed-back headphones. Responses were made via a custom keypad (P.I. Engineering, Williamston, MI).

2.1.4. Procedure.

A single-interval two-alternative forced choice task (1i-2afc) was used. On each trial, a single FM sweep was presented while a fixation cross and response choices (“Up” and “Down”) appeared on the screen. Half of trials contained an up sweep. The other half of trials contained down sweeps. Participants used the keypad to classify each sweep as “Up” or “Down”. After each response, there was a delay of 2 s plus jitter (uniformly distributed between 0 – 1 s) before the presentation of another FM sound. Participants completed 5 blocks of 50 trials each with a titration procedure that began at an easy level (rate of change or octaves/s = 1.50). After each response, accuracy feedback was presented on the screen (“Correct” or “Incorrect”). Each correct response was followed by a decrease in octaves/s by 0.1634 when the current rate of change was greater than 0.50, or by 0.0204 when the current rate was 0.50 or less. Each incorrect response was followed by an increase in octaves/s by 0.4000 when the current rate of change was greater than 0.50, or by 0.0500 when the current rate was 0.50 or less. We made step size larger when participants were classifying easier levels (rate of change > 0.50) to accelerate the titration procedure. Step sizes were smaller following correct responses to capture a more accurate threshold level. This targeted percent correct accuracy of ~71.7 [19].

2.2. Results.

Participants’ mean rate of change in pitch at the end of the thresholding task was 0.28 octaves/s (SD = 0.12). Average performance on the last 10 trials of the pilot experiment was 76.3% correct.

3. Main Experiment

Five stimulus levels for the main experiment were selected ranging from −2 to 2 SDs relative to the mean 0.28 rate of change determined by the Pilot Experiment. These levels were 0.04, 0.16, 0.28, 0.40, and 0.52 octaves/s.

3.1. Material and Methods

3.1.1. Participants.

Twenty-eight individuals (15 females and 13 males; ages 18 – 67, M = 28.43, SD = 14.04) participated for cash payment ($15 per hour) from KSU. Participants in the main experiment were recruited from KSU students, staff, and faculty. None of the participants in the main experiment participated in the pilot experiment. Because participants in the main study included faculty and staff, many were significantly older than participants in the pilot experiment which comprised undergraduate students. All participants in the main experiment were given a hearing screening using an Amplivox 170 audiometer. Only individuals with hearing thresholds <20 dB HL (.125 – 8 kHz) were allowed to continue participation. Two individuals were dropped from the sample because of excessive skin potentials in EEG data that would make it impossible to obtain a clear ERP waveform. One individual was dropped for excessively poor performance (>2 SD below mean). A total of 25 individuals were included in the final analyses of the main experiment.

3.1.2. Stimuli.

Same as in Pilot Experiment except now only five stimulus levels were presented.

3.1.3. Procedure.

The main experiment used the same 1i-2afc task with the stimulus levels noted above. The experiment started with 20 practice trials so that participants could get a sense of what “up” and “down” means in the context to the experiment. These trials all occurred with an easy FM rate (1.50 octaves/s) and were accompanied by feedback.

Three main differences in this task existed for the remainder of the experiment. First, the presented rate on each trial was sampled from the rates selected based on the pilot data. Second, participants were prompted to make confidence ratings after each trial. Following each “up” or “down” response, participants made a confidence rating about the accuracy of their response on a 6-point scale using the keypad (1 = “lowest confidence” to 6 = “highest confidence”). Third, no feedback was presented regarding correctness to prevent introducing bias to the confidence ratings. Participants completed 500 experiment trials, each presenting one of five FM rates for an up or down sweep (~100 per level). FM rates were ordered randomly within blocks. The entire task took approximately 40 minutes to complete, with short breaks between experimental blocks (50 trials each).

3.1.4. Equipment and Electrophysiology.

EEG was recorded using a Biosemi Active II system (Biosemi, Netherlands). Participants sat in a sound attenuating booth while hearing sounds over Etymotic ER-2 earphones (Etymotic, Research, Elk Grove Village, IL, USA) connected to an RME Fireface UC audio device. Responses were made via a custom keypad (P.I. Engineering, Williamston, MI).

Data were collected from a 68-channel array of electrodes at a 2048 Hz sampling rate and 24-bit A/D resolution. Data were referenced online to the Common-Mode-Sense/Driven-Right-Leg (CMS/DRL) reference of the BioSemi system (see www.biosemi.com). Sixty-four electrodes were fixed within a cap consistent with the international 10–20 placement system. Four additional electrodes were placed at the mastoids, and on lateral sides of each eye. Electrode offsets relative to CMS/DRL were brought within 25 μV or else were rejected from analysis.

All offline analyses were performed using EEGLAB ([20]; http://sccn.ucsd.edu/eeglab) and custom MATLAB scripts/functions. The data were referenced to the average of mastoids and resampled at 256 Hz (after applying a zero-phase antialiasing filter). An additional digital high-pass filter (1691-point zero phase FIR, 0.25-Hz half amplitude cutoff) and low-pass filter (69-point zero phase FIR, 56.25-Hz half-amplitude cutoff) were applied with the pop_eegfiltnew() function of the EEGLAB toolbox in order to remove slow-wave drifts and high-frequency activity that was not of interest for the planned ERP analyses. Based on visual inspection, sections of continuous data and channels contaminated by excessive noise or movement artifacts (that will make it impossible to obtain clear ERP waveforms) were manually removed. This includes high-frequency jaw clenching and low-frequency movement-related activity. The remaining data included, on average, 65 channels (min = 62, max = 66) and 423 trials (min = 211, max 491). Remaining data were submitted to infomax independent components analysis (ICA). Independent components (ICs) identified as artifacts (e.g., eye blinks) were subsequently removed from the channel data [21]. Excluded channels were interpolated. Epochs were extracted from −1000 to 3000 ms surrounding the tone onset. Mean baseline voltages were subtracted (−200 to 0 ms). ERPs were generated by averaging single-epoch voltage time courses.

3.2. Statistical Analyses.

For behavioral data, effects of stimulus level (rate of change .04, .16, .28, .40, or .52) on confidence and accuracy were analyzed with standard repeated-measures ANOVAs.

Amplitudes of ERP components were the mean voltage within time-windows determined from the grand-average waveforms across all participants. Time windows for each component following tone onset were selected based on a visual inspection of ERP waveforms and amplitudes across all electrodes (see Supplemental Materials). Average amplitudes were analyzed for the N1 component (140–180 ms) and P2 component (200 – 300 ms) at frontal and central clusters, separately. Average amplitudes were analyzed for the LPP component (500 – 900 ms) at the parietal cluster. The frontal cluster included electrodes AF3, AFz, AF4, F3, F1, Fz, F2, and F4. The central cluster included electrodes FC1, FCz, FC2, C1, Cz, C2, CP1, CPz, and CP2. The parietal cluster included electrodes P3, P1, Pz, P2, P4, PO3, POz, and PO4. Each cluster was selected based on previous work [15] for each component and the number of electrodes in the sensor array. Average amplitudes were calculated for each stimulus level on correct, incorrect, high confidence (correct only), and low confidence (correct only) trials.

Statistical analyses of ERPs were accomplished with separate linear mixed effects models fitted with maximum likelihood to predict component amplitude data with either correctness or confidence as a fixed factor. Stimulus level (centered and continuous) and the interaction were also treated as fixed factors. Subject intercepts and slopes for correctness or confidence, and stimulus level were treated as random factors. The significance of model terms was assessed by comparing the log likelihood of this full model with a reduced model with the component of interest removed. Matlab Statistics Toolbox (Mathworks, Natick, Massachusetts) was used for all statistical analyses.

3.3. Results

3.3.1. Behavioral Results.

With confidence as the dependent variable, we found a significant main effect of stimulus level, F(4,96) = 30.15, p < .001, η2p = .56. Mean confidence increased with rate of change, from 3.85 (SD = .92) (on the scale of 1–6) at 0.04 octaves/s to 4.72 (SD = .51) at 0.52. With proportion correct as the dependent variable, we found a significant main effect of stimulus level, F(4,96) = 178.57, p < .001, η2p = .88. Mean performance accuracy increased with rate of change, from 52% correct (SD = .04) at 0.04 octaves/s to 86% correct (SD = 11) at 0.52. These results were unsurprising as it has been well demonstrated that manipulations to stimulus difficulty impact accuracy and confidence in this manner [16,22–26]. See Figure 1AB for individual participants’ mean proportion correct and confidence by stimulus level.

Figure 1.

Figure 1.

Open in a new tab

Mean proportion correct (A) and confidence (B) at each stimulus level (rate of change). Each gray line represents a single participant. The thick black line is the mean across participants.

For a threshold-only analysis of ERP amplitude differences in confidence (section 3.3.2.2), the level that produced performance closest to 71% correct was selected for each participant (hereafter, “threshold level”). This was a rate of change of .04 for one participant, 0.16 for fourteen participants, 0.28 for eight participants, and 0.40 for two participants. Overall, these selected stimulus levels produced M = 71.7% correct performance (SD = 5.5%) when considered on their own.

3.3.2. ERP Results.

3.3.2.1. Weak relationships between accuracy and ERP amplitudes.

For N1 measured at central electrodes, there were no significant fixed effects or interaction, ps > .10. The same was true for the P2 linear mixed effects model, ps > .10.

For LPP, there was a significant effect of correctness, X2(1) = 4.30, p = .038, βcorrectness = −.59, 95% CI = [−1.14 −.05], related to a lower amplitude for incorrect compared to correct trials. There were no other significant effects, p > .10.

3.3.2.2. LPP amplitudes parallel confidence judgments.

Each participant’s mean confidence level was calculated across all trials in the experiment. High and low confidence trials were determined as trials above or below each participant’s mean confidence rating. Only correct trials were included in this analysis to control a potential confound in the impact of accuracy on amplitude differences. That is, a confidence difference cannot be well explained by the correctness effect reported above.

Neither the N1 nor P2 showed any significant effects, ps > .10. For LPP, there was a significant effect of confidence, X2(1) = 5.28, p = .022, βconfidence = −.54, 95% CI = [−.98 −.09], related to a lower amplitude for low confidence compared to high-confidence trials (see Figure 2B). There were no other significant effects, p > .10. This replicates previous work that revealed significant differences in stimulus-locked LPP amplitude for high compared to low confidence trials in an auditory detection task [15]. Mean voltage differences for high minus low confidence trials are shown in Figure 2B with standard error shading. The scalp topography of these mean differences during the LPP window 500–900 confirms this effect occurred at the parietal region.

Figure 2.

Figure 2.

Open in a new tab

Stimulus-locked ERPs at parietal electrodes (shown on left) for correct (solid), wrong (dotted), and correct minus wrong (green) trials (A) as well as for high confidence (solid), low confidence (dotted), and high minus low confidence (green), for correct-only trials (B). Green shading represents standard error. Scalp topography shows the regions of mean voltage differences during the LPP window (500 to 900 ms).

Mean LPP amplitude was also calculated for trials presenting participant’s threshold stimulus level (rate of change producing ~71.7% correct performance). This was done to eliminate the possibility that a stimulus effect (even if non-significant in the overall analysis) may have contributed to the LPP amplitude difference. The effect of confidence remained, X2(1) = 4.94, p = .026, βconfidence = −.1.18, 95% CI = [−2.20 −.17]. This replicates the main effect of confidence found across all stimulus levels. This finding suggests that, in this task, neural correlates of confidence did not vary based on difficulty level.

3.3.2.3. LPP is related to FM sweep processing, not response execution.

One potential concern with the LPP is that its latency could indicate that it is more closely related to the response than with the processing of the stimulus. To examine this, we conducted a response time sorted analyses where all epochs across conditions and participants were combined and sorted in respect to response time. An “erpimage” [20] was then made by averaging overlapping 60-trial windows at parietal electrodes. Figure 3 shows this image, with response time indicated by the blue line. It is clear from the figure that LPP latency is consistently locked to stimulus onset, regardless of single-trial response times. This is evident in the broad positive voltage deflection (red coloring in figure) starting at ~ 200 ms and lasting until ~1200 ms for both fast and slow response times.

Figure 3.

Figure 3.

Open in a new tab

Response-time sorted erpimage at parietal electrodes shown in Figure 2. Here, trial is on the y-axis, time is on the x-axis, and the color represents voltage for a moving average window across trials sorted by response time. The blue line represents mean response time across the moving windows. The white dotted line represents the LPP window (500 to 900 ms).

4. Discussion

The present study was designed to extend our recent work examining ERP correlates of metacognitive judgments in auditory detection [15] to an auditory categorization task. LPP amplitude paralleled confidence as in the previous study. Expanding on that work, this relationship was demonstrated to occur regardless of accuracy as it was present for correct-only trials. It also occurred independent of task difficulty as it was present for trials presenting stimuli at individually identified threshold levels. Finally, we confirmed that this LPP was not driven by response preparation after sorting ERP epochs by response time. Rather it likely reflects stimulus processing. Interestingly, there did not appear to be any relationship between amplitudes of the N1-P2 complex and confidence judgments. This was somewhat unexpected as we did observe this relationship previously in a detection task (also, see [22]).

We know that metacognitive judgments like confidence vary with performance and task difficulty [16,23–27]. Figure 1 illustrates this in our behavioral data. We also know that there is an association between perceptual discrimination difficulty and LPP amplitudes [28–31]. Unfortunately, much of the previous work on ERPs and confidence has collapsed across levels of difficulty and/or accuracy before examining relationships between confidence and ERPs (e.g., [32–34]). This has left the relationship somewhat tenuous. One cannot be sure as to whether or not many effects may have been driven by task difficulty or accuracy rather than a relationship between confidence and the ERP. The current data importantly demonstrate that ERP differences occur between confidence levels on trials with similar difficulty and accuracy (also see [9,12,36]). One caveat is that the present study’s design and data do not allow for a full dissociation of confidence and task performance, for example, keeping threshold stimuli constant on a trial-by-trial basis (for discussion, see [37]). Nevertheless, this auditory categorization work, combined with other findings that the LPP, P3, and/or late positive component (LPC) parallels confidence in a variety of tasks (e.g., auditory detection – [15]; visual perception – [38]; episodic memory – [39]; recollection – [10]) suggests that the LPP is a task general indication of the confidence one will have for an upcoming judgment. This would make it worthwhile to use LPPs in a number of ways from covert monitoring of confidence in studies (e.g., without having to query confidence on every trial), to input to classifiers aimed at decoding confidence from individuals’ behaviors and physiological data (e.g., [35,38,32,36]). Even though our findings did not allow for a complete attribution of the LPP difference to performance-independent confidence per se, the LPP amplitude might retain utility in such machine learning research.

Why did we not observe a relationship between amplitudes of the N1 and P2 of the auditory ERP, while others have? Though we did not explicitly manipulate task complexity here, we suspect that whereas the LPP is a task general indicator of confidence, earlier “evoked” components of the ERP might be task specific. Previous studies found the amplitude of N1 and P2 varied with confidence when listeners were tasked to detect sounds in noise [22,15]. In detection tasks, these earlier responses are indicative that the brain has registered the onset of a stimulus and may also be related to awareness that a target sound occurred. For instance, Gutschalk et al. [40] found that awareness of a repeating tone in the presence of masking built up over the course of repeats and paralleled an increasing amplitude for a negative going component peaking at around 100 ms post-tone onset. Interestingly, this component was not present on trials where the repeating tone went undetected. In our experiment, all sounds were presented well above detection threshold. Correct responding (and hence confidence) may be related more to later processing stages in categorization tasks like this one. Future work will need to determine how the cognitive demands of a task are related to the stages of internal stimulus processing and the making of confidence judgments.

5. Conclusions

The current study along with the growing number of other studies measuring confidence judgments in a variety of different types of paradigms suggests that LPP is a general indicator of confidence. The relationship between confidence and earlier ERP features may depend on the cognitive demands of a given task.

Supplementary Material

1

  • Tone-locked late posterior positivity amplitudes increased with confidence level

  • Neural correlates of confidence did not vary based on difficulty level

  • The relationship between confidence and N1 and P2 may depend on task demands

Acknowledgements

This work was supported by the Cognitive and Neurobiological Approaches to Plasticity (CNAP) Center of Biomedical Research Excellence (COBRE) of the National Institutes of Health under grant number P20GM113109. The content is solely the responsibility of the authors and does not necessarily represent the official views of the National Institutes of Health.

Footnotes

Publisher's Disclaimer: This is a PDF file of an unedited manuscript that has been accepted for publication. As a service to our customers we are providing this early version of the manuscript. The manuscript will undergo copyediting, typesetting, and review of the resulting proof before it is published in its final form. Please note that during the production process errors may be discovered which could affect the content, and all legal disclaimers that apply to the journal pertain.

Declaration of Competing Interest

The authors declare that they have no known competing financial interests or personal relationships that could have appeared to influence the work reported in this paper.

CRediT authorship contribution statement

A.C. Zakrzewski: Investigation, Formal analysis, Writing – Original Draft, Funding acquisition, Conceptualization, Methodology; B. Maniscalco: Supervision, Writing – Review & Editing, Conceptualization, Methodology; M.G. Wisniewski: Resources, Formal analysis, Writing – Review & Editing, Conceptualization, Methodology

Data availability

Data will be made available on request.

References

  • [1].Desender K, Boldt A, Yeung N, Subjective Confidence Predicts Information Seeking in Decision Making, Psychol Sci 29 (2018) 761–778. [DOI] [PubMed] [Google Scholar]
  • [2].Lachman JL, Lachman R, Thronesbery C, Metamemory through the adult life span, Developmental Psychology 15 (1979) 543. [Google Scholar]
  • [3].Baird B, Smallwood J, Gorgolewski KJ, Margulies DS, Medial and lateral networks in anterior prefrontal cortex support metacognitive ability for memory and perception, Journal of Neuroscience 33 (2013) 16657–16665. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • [4].Baird B, Cieslak M, Smallwood J, Grafton ST, Schooler JW, Regional white matter variation associated with domain-specific metacognitive accuracy, Journal of cognitive neuroscience 27 (2015) 440–452. [DOI] [PubMed] [Google Scholar]
  • [5].McCurdy LY, Maniscalco B, Metcalfe J, Liu KY, De Lange FP, Lau H, Anatomical coupling between distinct metacognitive systems for memory and visual perception, Journal of Neuroscience 33 (2013) 1897–1906. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • [6].Fleming SM, Ryu J, Golfinos JG, Blackmon KE, Domain-specific impairment in metacognitive accuracy following anterior prefrontal lesions, Brain 137 (2014) 2811–2822. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • [7].Lee AL, Ruby E, Giles N, Lau H, Cross-domain association in metacognitive efficiency depends on first-order task types, Frontiers in psychology 9 (2018) 2464. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • [8].Morales J, Lau H, Fleming SM, Domain-general and domain-specific patterns of activity supporting metacognition in human prefrontal cortex, Journal of Neuroscience 38 (2018) 3534–3546. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • [9].Boldt A, Yeung N, Shared neural markers of decision confidence and error detection, Journal of Neuroscience 35 (2015) 3478–3484. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • [10].Curran T, Effects of attention and confidence on the hypothesized ERP correlates of recollection and familiarity, Neuropsychologia 42 (2004) 1088–1106. [DOI] [PubMed] [Google Scholar]
  • [11].Selimbeyoglu A, Keskin-Ergen Y, Demiralp T, What if you are not sure? Electroencephalographic correlates of subjective confidence level about a decision, Clinical Neurophysiology 123 (2012) 1158–1167. [DOI] [PubMed] [Google Scholar]
  • [12].Faivre N, Filevich E, Solovey G, Kühn S, Blanke O, Behavioral, modeling, and electrophysiological evidence for supramodality in human metacognition, Journal of Neuroscience 38 (2018) 263–277. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • [13].Masset P, Ott T, Lak A, Hirokawa J, Kepecs A, Behavior-and modality-general representation of confidence in orbitofrontal cortex, Cell 182 (2020) 112–126. e118. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • [14].Squires KC, Hillyard SA, Lindsay PH, Vertex potentials evoked during auditory signal detection: Relation to decision criteria, Perception & Psychophysics 14 (1973) 265–272. [Google Scholar]
  • [15].Zakrzewski AC, Wisniewski MG, Iyer N, Simpson BD, Confidence tracks sensory-and decision-related ERP dynamics during auditory detection, Brain and Cognition 129 (2019) 49–58. [DOI] [PubMed] [Google Scholar]
  • [16].Fleming SM, Dolan RJ, The neural basis of metacognitive ability, Philosophical Transactions of the Royal Society B: Biological Sciences 367 (2012) 1338–1349. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • [17].Rahnev D, Visual metacognition: Measures, models, and neural correlates, American Psychologist 76 (2021) 1445. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • [18].Wisniewski MG, Liu EH, Church BA, Mercado EM, Learning to discriminate frequency modulation rate can benefit and worsen pitch acuity, JASA Express Letters 135 (2014) 55–60. [DOI] [PubMed] [Google Scholar]
  • [19].Leek MR, Adaptive procedures in psychophysical research, Perception & psychophysics 63 (2001) 1279–1292. [DOI] [PubMed] [Google Scholar]
  • [20].Delorme A, Makeig S, EEGLAB: an open source toolbox for analysis of single-trial EEG dynamics including independent component analysis, Journal of neuroscience methods 134 (2004) 9–21. [DOI] [PubMed] [Google Scholar]
  • [21].Jung T-P, Makeig S, Westerfield M, Townsend J, Courchesne E, Sejnowski TJ, Removal of eye activity artifacts from visual event-related potentials in normal and clinical subjects, Clinical Neurophysiology 111 (2000) 1745–1758. [DOI] [PubMed] [Google Scholar]
  • [22].Hillyard SA, Squires KC, Bauer JW, Lindsay PH, Evoked potential correlates of auditory signal detection, Science 172 (1971) 1357–1360. [DOI] [PubMed] [Google Scholar]
  • [23].Kao Y-C, Davis ES, Gabrieli JD, Neural correlates of actual and predicted memory formation, Nature neuroscience 8 (2005) 1776–1783. [DOI] [PubMed] [Google Scholar]
  • [24].Kruger J, Dunning D, Unskilled and unaware of it: how difficulties in recognizing one’s own incompetence lead to inflated self-assessments, Journal of personality and social psychology 77 (1999) 1121. [DOI] [PubMed] [Google Scholar]
  • [25].Kunimoto C, Miller J, Pashler H, Confidence and accuracy of near-threshold discrimination responses, Consciousness and cognition 10 (2001) 294–340. [DOI] [PubMed] [Google Scholar]
  • [26].Modirrousta M, Fellows LK, Medial prefrontal cortex plays a critical and selective role in ‘feeling of knowing’meta-memory judgments, Neuropsychologia 46 (2008) 2958–2965. [DOI] [PubMed] [Google Scholar]
  • [27].Morgan MJ, Mason A, Solomon J, Blindsight in normal subjects, Nature 385 (1997) 401–402. [DOI] [PubMed] [Google Scholar]
  • [28].Ball NJ, Mercado III E, Orduña I, Enriched environments as a potential treatment for developmental disorders: a critical assessment, Frontiers in psychology 10 (2019) 466. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • [29].Hagen GF, Gatherwright JR, Lopez BA, Polich J, P3a from visual stimuli: task difficulty effects, International journal of psychophysiology 59 (2006) 8–14. [DOI] [PubMed] [Google Scholar]
  • [30].Squires KC, Squires NK, Hillyard SA, Decision-related cortical potentials during an auditory signal detection task with cued observation intervals, Journal of Experimental Psychology: Human Perception and Performance 1 (1975) 268. [DOI] [PubMed] [Google Scholar]
  • [31].Sutton S, Ruchkin DS, Munson R, Kietzman ML, Hammer M, Event-related potentials in a two-interval forced-choice detection task, Perception & Psychophysics 32 (1982) 360–374. [DOI] [PubMed] [Google Scholar]
  • [32].Gherman S, Philiastides MG, Neural representations of confidence emerge from the process of decision formation during perceptual choices, Neuroimage 106 (2015) 134–143. [DOI] [PubMed] [Google Scholar]
  • [33].Paller KA, Kutas M, Mayes AR, Neural correlates of encoding in an incidental learning paradigm, Electroencephalography and clinical neurophysiology 67 (1987) 360–371. [DOI] [PubMed] [Google Scholar]
  • [34].Selimbeyoglu A, Keskin-Ergen Y, Demiralp T, What if you are not sure? Electroencephalographic correlates of subjective confidence level about a decision, Clinical Neurophysiology 123 (2012) 1158–1167. [DOI] [PubMed] [Google Scholar]
  • [35].Boldt A, Yeung N, Shared neural markers of decision confidence and error detection, Journal of Neuroscience 35 (2015) 3478–3484. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • [36].Gherman S, Philiastides MG, Human VMPFC encodes early signatures of confidence in perceptual decisions, Elife 7 (2018) e38293. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • [37].Morales J, Chiang J, Lau H, Controlling for performance capacity confounds in neuroimaging studies of conscious awareness, Neuroscience of Consciousness 2015 (2015). [DOI] [PMC free article] [PubMed] [Google Scholar]
  • [38].Desender K, Murphy P, Boldt A, Verguts T, Yeung N, A postdecisional neural marker of confidence predicts Information-Seeking in Decision-Making, Journal of Neuroscience 39 (2019) 3309–3319. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • [39].Wynn SC, Kessels RP, Schutter DJ, Electrocortical indices of subjectively perceived confidence in episodic memory, International Journal of Psychophysiology 151 (2020) 18–24. [DOI] [PubMed] [Google Scholar]
  • [40].Gutschalk A, Micheyl C, Oxenham AJ, Neural correlates of auditory perceptual awareness under informational masking, PLoS biology 6 (2008) e138. [DOI] [PMC free article] [PubMed] [Google Scholar]

Associated Data

This section collects any data citations, data availability statements, or supplementary materials included in this article.

Supplementary Materials

1
NIHMS1900857-supplement-1.pdf (808.9KB, pdf)

Data Availability Statement

Data will be made available on request.

RESOURCES