The medial frontal cortex enables performance monitoring, indexed by the error-related negativity (ERN) and manifested by performance adaptations. We recorded electroencephalogram over and neural spiking across all layers of the supplementary eye field, an agranular cortical area, in monkeys performing a saccade-countermanding (stop signal) task. Neurons signaling error production, feedback predicting reward gain or loss, and delivery of fluid reward had different spike widths and were concentrated differently across layers. Neurons signaling error or loss of reward were more common in layers 2 and 3 (L2/3), whereas neurons signaling gain of reward were more common in layers 5 and 6 (L5/6). Variation of error– and reinforcement-related spike rates in L2/3 but not L5/6 predicted response time adaptation. Variation in error-related spike rate in L2/3 but not L5/6 predicted ERN magnitude. These findings reveal novel features of cortical microcircuitry supporting performance monitoring and confirm one cortical source of the ERN.
This is a preview of subscription content, access via your institution
Open Access articles citing this article.
Nature Communications Open Access 17 July 2020
Subscribe to Nature+
Get immediate online access to the entire Nature family of 50+ journals
Subscribe to Journal
Get full journal access for 1 year
only $8.25 per issue
All prices are NET prices.
VAT will be added later in the checkout.
Tax calculation will be finalised during checkout.
Get time limited or full article access on ReadCube.
All prices are NET prices.
The analysis codes that were used for this study are available from the corresponding author upon reasonable request.
The data that support the findings of this study are available from the corresponding author upon reasonable request.
Verbruggen, F. & Logan, G. D. Models of response inhibition in the stop-signal and stop-change paradigms. Neurosci. Biobehav. Rev. 33, 647–661 (2009).
Emeric, E. E. et al. Influence of history on saccade countermanding performance in humans and macaque monkeys. Vision Res. 47, 35–49 (2007).
Kolling, N. et al. Value, search, persistence and model updating in anterior cingulate cortex. Nat. Neurosci. 19, 1280–1285 (2016).
Shenhav, A., Cohen, J. D. & Botvinick, M. M. Dorsal anterior cingulate cortex and the value of control. Nat. Neurosci. 19, 1286–1291 (2016).
Gehring, W. J., Liu, Y., Orr, J. M. & Carp, J. The error-related negativity (ERN/Ne). In Oxford Handbook of Event-Related Potential Components (Luck, S. J. & Kappenman, E. eds.), 231–291 (Oxford University Press, New York, 2012).
Godlove, D. C. et al. Event-related potentials elicited by errors during the stop-signal task. I. Macaque monkeys. J. Neurosci. 31, 15640–15649 (2011).
Cohen, M. X. A neural microcircuit for cognitive conflict detection and signaling. Trends Neurosci. 37, 480–490 (2014).
Ito, S., Stuphorn, V., Brown, J. W. & Schall, J. D. Performance monitoring by the anterior cingulate cortex during saccade countermanding. Science 302, 120–122 (2003).
Emeric, E. E. et al. Performance monitoring local field potentials in the medial frontal cortex of primates: anterior cingulate cortex. J. Neurophysiol. 99, 759–772 (2008).
Cai, X. & Padoa-Schioppa, C. Neuronal encoding of subjective value in dorsal and ventral anterior cingulate cortex. J. Neurosci. 32, 3791–3808 (2012).
Monosov, I. E. Anterior cingulate is a source of valence-specific information about value and uncertainty. Nat. Commun. 8, 134 (2017).
Kawai, T., Yamada, H., Sato, N., Takada, M. & Matsumoto, M. Preferential representation of past outcome information and future choice behavior by putative inhibitory interneurons rather than putative pyramidal neurons in the primate dorsal anterior cingulate cortex. Cereb. Cortex https://dx.doi.org/10.1093/cercor/bhy103 (2018).
Stuphorn, V., Taylor, T. L. & Schall, J. D. Performance monitoring by the supplementary eye field. Nature 408, 857–860 (2000).
Emeric, E. E., Leslie, M., Pouget, P. & Schall, J. D. Performance monitoring local field potentials in the medial frontal cortex of primates: supplementary eye field. J. Neurophysiol. 104, 1523–1537 (2010).
Chen, X. & Stuphorn, V. Sequential selection of economic good and action in medial frontal cortex of macaques during value-based decisions. eLife 4, e09418 (2015).
Amiez, C. & Petrides, M. Anatomical organization of the eye fields in the human and non-human primate frontal cortex. Prog. Neurobiol. 89, 220–230 (2009).
Bonini, F. et al. Action monitoring and medial frontal cortex: Leading role of supplementary motor area. Science 343, 888–891 (2014).
Stuphorn, V., Brown, J. W. & Schall, J. D. Role of supplementary eye field in saccade initiation: Executive, not direct, control. J. Neurophysiol. 103, 801–816 (2010).
Stuphorn, V. & Schall, J. D. Executive control of countermanding saccades by the supplementary eye field. Nat. Neurosci. 9, 925–931 (2006).
Bastos, A. M. et al. Canonical microcircuits for predictive coding. Neuron 76, 695–711 (2012).
Shipp, S. The importance of being agranular: a comparative account of visual and motor cortex. Phil. Trans. R. Soc. Lond. B 360, 797–814 (2005).
Godlove, D. C., Maier, A., Woodman, G. F. & Schall, J. D. Microcircuitry of agranular frontal cortex: Testing the generality of the canonical cortical microcircuit. J. Neurosci. 34, 5355–5369 (2014).
Beul, S. F. & Hilgetag, C. C. Towards a “canonical” agranular cortical microcircuit. Front. Neuroanat. 8, 165 (2015).
Ninomiya, T., Dougherty, K., Godlove, D. C., Schall, J. D. & Maier, A. Microcircuitry of agranular frontal cortex: contrasting laminar connectivity between occipital and frontal areas. J. Neurophysiol. 113, 3242–3255 (2015).
Hanes, D. P. & Schall, J. D. Countermanding saccades in macaque. Vis. Neurosci. 12, 929–937 (1995).
Kawaguchi, N. et al. Surprise signals in the supplementary eye field: rectified prediction errors drive exploration-exploitation transitions. J. Neurophysiol. 113, 1001–1014 (2015).
Vigneswaran, G., Kraskov, A. & Lemon, R. N. Large identified pyramidal cells in macaque motor and premotor cortex exhibit “thin spikes”: Implications for cell type classification. J. Neurosci. 31, 14235–14242 (2011).
Stahl, J. & Gibbons, H. Dynamics of response-conflict monitoring and individual differences in response control and behavioral control: an electrophysiological investigation using a stop-signal task. Clin. Neurophysiol. 118, 581–596 (2007).
Chang, A., Chen, C. C., Li, H. H. & Li, C. S. Event-related potentials for post-error and post-conflict slowing. PLoS One 9, e99909 (2014).
Reinhart, R. M., Carlisle, N. B., Kang, M. S. & Woodman, G. F. Event-related potentials elicited by errors during the stop-signal task. II: human effector-specific error responses. J. Neurophysiol. 107, 2794–2807 (2012).
Cohen, M. X. Where does EEG come from and what does it mean? Trends Neurosci. 40, 208–218 (2017).
Riera, J. J. et al. Pitfalls in the dipolar model for the neocortical EEG sources. J. Neurophysiol. 108, 956–975 (2012).
Schultz, W. Multiple reward signals in the brain. Nat. Rev. Neurosci. 1, 199–207 (2000).
Matsumoto, M. & Hikosaka, O. Representation of negative motivational value in the primate lateral habenula. Nat. Neurosci. 12, 77–84 (2009).
Huerta, M. F. & Kaas, J. H. Supplementary eye field as defined by intracortical microstimulation: Connections in macaques. J. Comp. Neurol. 293, 299–330 (1990).
Parthasarathy, H. B., Schall, J. D. & Graybiel, A. M. Distributed but convergent ordering of corticostriatal projections: Analysis of the frontal eye field and the supplementary eye field in the macaque monkey. J. Neurosci. 12, 4468–4488 (1992).
Williams, S. M. & Goldman-Rakic, P. S. Widespread origin of the primate mesofrontal dopamine system. Cereb. Cortex 8, 321–345 (1998).
Aston-Jones, G. & Cohen, J. D. Adaptive gain and the role of the locus coeruleus-norepinephrine system in optimal performance. J. Comp. Neurol. 493, 99–110 (2005).
Giguere, M. & Goldman-Rakic, P. S. Mediodorsal nucleus: areal, laminar, and tangential distribution of afferents and efferents in the frontal lobe of rhesus monkeys. J. Comp. Neurol. 277, 195–213 (1988).
Sommer, M. A. & Wurtz, R. H. Brain circuits for the internal monitoring of movements. Annu. Rev. Neurosci. 31, 317–338 (2008).
Thakkar, K. N., Schall, J. D., Boucher, L., Logan, G. D. & Park, S. Response inhibition and response monitoring in a saccadic countermanding task in schizophrenia. Biol. Psychiatry 69, 55–62 (2011).
Foti, D., Kotov, R., Bromet, E. & Hajcak, G. Beyond the broken error-related negativity: Functional and diagnostic correlates of error processing in psychosis. Biol. Psychiatry 71, 864–872 (2012).
Thakkar, K. N., Diwadkar, V. A. & Rolfs, M. Oculomotor prediction: A window into the psychotic mind. Trends. Cogn. Sci. 21, 344–356 (2017).
Pouget, P. et al. Neural basis of adaptive response time adjustment during saccade countermanding. J. Neurosci. 31, 12604–12612 (2011).
Kravitz, A. V., Tye, L. D. & Kreitzer, A. C. Distinct roles for direct and indirect pathway striatal neurons in reinforcement. Nat. Neurosci. 15, 816–818 (2012).
Vicente, A. M., Galvão-Ferreira, P., Tecuapetla, F. & Costa, R. M. Direct and indirect dorsolateral striatum pathways reinforce different action strategies. Curr. Biol. 26, R267–R269 (2016).
Boucher, L., Palmeri, T. J., Logan, G. D. & Schall, J. D. Inhibitory control in mind and brain: an interactive race model of countermanding saccades. Psychol. Rev. 114, 376–397 (2007).
Hubel, D. H. & Wiesel, T. N. Receptive fields and functional architecture of monkey striate cortex. J. Physiol. (Lond.) 195, 215–243 (1968).
Chandrasekaran, C., Peixoto, D., Newsome, W. T. & Shenoy, K. V. Laminar differences in decision-related neural activity in dorsal premotor cortex. Nat. Commun. 8, 614 (2017).
Bastos, A. M., Loonis, R., Kornblith, S., Lundqvist, M. & Miller, E. K. Laminar recordings in frontal cortex suggest distinct layers for maintenance and control of working memory. Proc. Natl Acad. Sci. USA 115, 1117–1122 (2018).
The authors thank J. Elsey M. Feurtado, M. Maddox, S. Motorny, J. Parker, M. Schall, C.R. Subraveti, and L. Toy for animal care and other technical assistance, J. Kaas for helpful comments and sharing histological material, and J. Brown, S. Errington, A. Maier, T. Reppert, M. Servant, V. Stuphorn, A. Tomarken, T. Womelsdorf, G. Woodman, and the reviewers for helpful discussions and comments on the manuscript. Imaging data was collected in the Vanderbilt Institute of Imaging Science. This work was supported by R01-MH55806 (JS), P30-EY08126 (JS), and by Robin and Richard Patton through the E. Bronson Ingram Chair in Neuroscience (JS).
The authors declare no competing interests.
Publisher’s note: Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.
Integrated supplementary information
a, Results of an automated procedure for aligning linear electrode array recordings across sessions based on current source density (CSD) sinks derived from visually-evoked potential (VEP). Top panels show visually evoked CSD recorded during each session masked to show locations of the 4 grand-averaged visually evoked sinks, with monkey and location of each penetration indicated above. Note the close correspondence in location of these sinks across recording sessions demonstrating the similarity in CSD recorded across sessions and success of the alignment procedure. Horizontal black bars indicate the estimated average location of gray matter based on the physiological signals. Bottom panels portray physiological signals observed on individual sessions. Electrocardiogram (red lines), elevated gamma activity (blue lines), and isolated single-units (black triangles) correspond with the estimate of the location of gray matter (gray shading). Yellow arrow highlights the plot from the session shown in Figure 1. Modified from22. b, Results of an automated procedure for aligning linear electrode array recordings across sessions based on LFP frequency content. Conventions as in a Modified from24. c, Distribution of depths of all neurons sampled during the saccade countermanding task from the sessions with penetrations perpendicular to the cortical layers. The color time-depth plots in Figure 2 and Figure 4 are obtained by correcting the number of neurons at each depth against this distribution. d, Consistency of laminar unit recordings within and across sites and monkeys. Normalized difference in discharge rate between error noncanceled trials and correct no stop signal trials at each depth were converted to a color map ranging from facilitation (>0) to suppression (<0), scaled across all depths. Left two panels illustrate post-saccadic modulation recorded at site P1 in one monkey on two sessions. Right two panels illustrate modulation with fluid delivery recorded in different monkeys. The consistency of the laminar variation of neural modulation demonstrates the existence of a functional architecture. The reproducibility of neuron types across independent measures across both monkeys is detailed in Supplementary Table 1. e, Laminar structure of SEF. From left to right are shown sections through SEF stained for Nissl, NeuN, nonphosphorylated neurofilament H (SMI-32), cytochrome oxidase (CO), Gallyas myelin, vesicular glutamate transporter 2 (VGLUT2), acetylcholinesterase (AChE), calretinin (CR), calbindin (CB), and parvalbumin (PV). Counts of CR, CB, and PV stained neurons are shown from22. Scale bar indicates 1 mm. In SEF, L1 is ~200 µm thick with some CR and CB but no PV neurons, and weak staining for CO, myelin, and VGLUT2. L2 is ~300 µm thick with dense stellate and few pyramidal neurons, the highest density of CR and CB neurons, but no SMI-32, stronger CO staining, fascicular myelin fibers, slightly stronger VGLUT2, and modest AChE. L3 is ~700 µm thick with a superficial sublayer with smaller pyramidal neurons and weak SMI-32 staining, very few CR, and modest densties of CB and PV neurons, stronger CO staining and denser myelin, VGLUT2, and AChE. A deeper sublayer is characterized by larger pyramidal neurons, pronounced SMI-32 staining, vanishingly few CR, less dense CB and modestly dense PV neurons, weaker CO, denser myelin and VGLUT2, and denser AChE. No granular L4 is evident in SEF. L5 is ~300 µm thick with large pyramidal neurons but inconsistent SMI-32 staining, no CR and fewer CB but modest density of PV neurons, lighter CO, denser myelin, lighter VGLUT2, and diminishing AChE staining. L6 is ~700 µm thick with smaller pyramidal neurons, light SMI-32, no CR, vanishingly few CB and low density of PV neurons, with lighter CO, still denser myelin, lighter VGLUT2 and sparse AChE staining. This variation in laminar structure guides the investigation of functional architecture of SEF.
a, Raster and spike density plots for the representative Error neuron aligned on target presentation (leftmost), stop signal (middle left), saccade (middle right), and feedback tone (rightmost) for error (top row rasters; thick dotted line) and latency-matched no stop signal trials (middle row rasters; thin solid line) trials as well as canceled trials (bottom row rasters; thick solid line). b, Spike density plots from the representative neuron for correct (solid) trials and subset of error trials (dashed) with no refixation saccade (n = 5 trials). On sessions in which the monkey often made refixation saccades after the error, exclusion of these trials resulted in very few trials (often less than 10 trials) for meaningful interpretation of the spike density function. Therefore, we devised an alternative analysis to ensure that the modulation is not simply saccade-related activity. We characterized neurons as error-related only if the beginning of modulation did not covary with the latency of the return-to-fixation saccade, but was synchronized with the error movement. c, Illustration of the approach using data for the representative neuron. Horizontal eye position (top) and spike density function (bottom) for error trials with fast (blue) and slow (red) return-to-fixation saccades. The onset of modulation for this Error neuron does not depend on the latency of the return-to-fixation saccade. Of the sample of 132 putative error neurons, 71 neurons showed a shift in modulation time that covaried with the variation in timing of the refixation saccade and thus were rejected. The remaining 61 Error neurons included in other analyses exhibited modulation times that were more closely locked to the time of the error saccade rather than the time of the refixation saccade. d, Population activity for the 61 included Error neurons (top row) and 71 rejected Error neurons (bottom row) based on the criterion described above. Activity was not normalized and is aligned on the error saccade (left) and refixation saccade (right) for no stop signal trials (black), fast refixation saccades (blue), and slow refixation saccades (red). Standard errors around the mean are shaded. For the population of Error neurons included in subsequent analyses, the onset of modulation on noncanceled trials was invariant relative to the timing of the refixation saccade. In contrast, for the population of neurons rejected for further analysis, the timing of modulation varied with the latency of the refixation saccade.
To examine whether Error, Loss, and Gain neurons respond to events in other task epochs associated with a loss of opportunity for gaining reward, we contrasted the task-related modulation of each neuron type against that measured after fixation breaks before the target was presented, which prematurely terminated the trial. Note that fixation breaks occur when monkeys are distracted or unmotivated, so this behavior is not equivalent to task-related errors in which monkeys attempt but fail to obtain reward. Also, because fixation breaks occur shortly after achieving central fixation, without much investment in the trial, they may not engage the same error and reinforcement mechanisms. Nevertheless, to gain further insight into the nature of the various signal types reported in this study, we compared the valence of modulation for fixation breaks and task-related errors by defining two contrasts: Modulation Indexfixation break = [Activity(no break fixation) – Activity(break fixation)] / [Activity(no break fixation) + Activity(break fixation)] and Modulation Indextask-related = [Activity(correct or reward) – Activity(error or loss)] / [Activity(correct or reward) + Activity(error or loss)]. Modulation related to fixation breaks was measured within 400 ms of trial onset, during which the vast majority of fixation breaks occurred. This was contrasted against trials with no fixation break. For Error neurons, the task-related activity corresponded to the post-saccadic period when activity was significantly different on error compared to correct trials. For Gain and Loss neurons, it corresponded to the post-tone period when activity was significantly different on unrewarded compared to rewarded trials. a, Scatterplots of the modulation measured on fixation breaks as a function of that measured on error/loss trials. Open circles mark neurons for which Modulation Indexbreak fixation was not significantly different from 0.0. Closed circles mark neurons for which Modulation Indexbreak fixation was significantly different from 0.0. The left plot shows the Error neurons. The right plot includes Loss and Gain neurons. By definition, Error and Loss neurons have negative Modulation indextask-related and Gain neurons have positive Modulation indextask-related. Note that the majority of neurons, irrespective of functional category, have negative Modulation Indexbreak fixation indicating higher activity upon fixation breaks. b, Tabular summary of the plotted values. A statistically significant majority of modulated neurons of each functional type showed higher activity in relation to fixation breaks (that is, negative Modulation Indexbreak fixation; Error: χ2(1, 34) = 13.9, p < 0.001; Loss: χ2(1, 130) = 9.98, p = 0.0016; Gain: χ2(1, 62) = 6.66, p = 0.0098). Relative to Gain and Loss neurons, a significantly higher proportion of Error neurons had higher activity associated with fixation breaks (Error vs. Loss: χ2(1, 166) = 6.94, p = 0.0084; Error vs. Gain: χ2(1, 97) = 4.86, p = 0.027). One might expect that Gain neurons would trend opposite the Loss neurons, but the proportions of Gain and Loss neurons with negative Modulation Index are indistinguishable (χ2(1, 193) = 0.20, p = 0.66). The fact that each type of neuron was similarly modulated in relation to fixation breaks suggests that the modulation at the time of fixation-breaks may be under the influence of the state of the monkey before the trial began rather than simply information about anticipated gain or loss of reward, as observed during the task-relevant period.
a, Distribution of spike widths across the sample reveals bimodal distribution. Dashed line marks 250 µs separation. b, Variation of spike widths across depth for neurons sampled in perpendicular penetrations (n = 293). The number of neurons at each time-depth indicated by gray scale (black = 5 neurons). The width of the spikes narrower than 250 µs does not vary with depth, and the incidence of encountering narrow spikes parallels the density of PV neurons. The width of spikes wider than 250 µs increases from L3 to L6, which parallels the size of pyramidal neurons. Also, the incidence of isolated neurons decreases with depth, which parallels the density of pyramidal neurons. c, Cumulative distribution of spike width for Error neurons in the entire sample (n = 61, solid) and those recorded in penetrations oriented perpendicular to the cortical layers (n = 42, dotted). The overlapping distributions is indicative of sampling from the same population.
a, Grand average ERP voltage on correct no stop signal trials (narrow solid) and error noncancelled stop signal trials (thick dotted) for monkeys Eu (left) and X (right). Saccade spike potential is prominent in both types of trials, but polarization is significantly more negative within an early period (~100–250 ms) following errors, characteristic of the ERN. b, Heat-map event-related potential raster for error (left) and correct (right) trials during a representative session. EEG voltage magnitudes on ~270 noncanceled stop signal trials (left) and ~470 no stop signal trials are ordered from top to bottom by the voltage at the time of peak ERN (ERNpeak at 187ms). The blue evident in the heat plot across trials indicates that sampling voltage at this time provides a reliable measure of polarization negativity. c, Distribution of the measure of EEG polarization pooled across error and correct trials. Values approximated a normal distribution. Outliers exceeding 3 standard deviations (dashed vertical lines) from the mean (solid vertical line) were excluded. Outliers were very rare, as indicated. Sampling the EEG in this way allowed discriminability between error and correct trials as indicated by the ROC curve constructed from the distributions of ERN on correct and error trials. The area under the ROC curve (AUC) was 0.61. d, Proportion of error trials (left) and the average over time of EEG voltage (right) for each bin. Other methods of ERN sampling were also considered in analyses related to ERN-spike relationship. See Supplementary Table 2 for summary statistics.
a, Zero-order Spearman’s rank correlation for relationship between EEG magnitude and spiking activity of Error (left) and non-error (right) neurons sampled in L2/3 (top) and L5/6 (bottom), complementing Figure 3. Per EEG convention, negative values are plotted above on the ordinate scale. Normalized EEG polarization magnitude was correlated with normalized spike rates of both Error and non-error neurons in both L2/3 and L5/6. The associated rS and p values are written in black above each panel. Statistically significant relationships are highlighted by lines of best fit. However, after accounting for (1) the correlation of spike rates across L2/3 and L5/6 of both Error and non-error neurons and (2) correct versus error trial outcome, only the variation of spike rate of Error neurons in L2/3 predicted variation of EEG polarization. The associated rS and p values for the partial correlation are written in red above each panel. b, Zero-order Spearman’s rank correlation for relationship between EEG magnitude and spiking activity of Error neurons on correct (left) and error (right) trials separately sampled in L2/3 (top) and L5/6 (bottom). Normalized EEG polarization magnitude was correlated with normalized Aerror in L2/3 on both correct and error trials but only in L5/6 in error trials. The associated rS and p values are written in black above each panel. Statistically significant relationships are highlighted by lines of best fit. c, Partial correlation plots corresponding to panels in b. Along the ordinate is plotted, using inverted EEG convention, residual fixed-effects-adjusted EEG voltage rank controlling for the ranks of fixed-effects-adjusted activity in the opposite layer. Along the abscissa is plotted the residual fixed-effects adjusted Aerror rank in the identified layer controlling for the fixed-effects adjusted activity in the opposite layer. The associated rS and p values for the partial correlation are written in red above each panel. Statistically significant relationships are highlighted by lines of best fit. In each panel, each point plots the average EEG voltage and associated spike count in one of 20 bins, ranked by voltage magnitude, with equal numbers of trials per bin. Thus, the number of points plotted equals 20 per session for the 6 sessions with non-zero spike counts in both L2/3 and L5/6. Summary statistics for other methods of measuring ERN magnitude are presented in Supplementary Table 2.
a, Raster and spike density plots for the representative Gain neuron aligned on target presentation (leftmost), stop signal (middle left), saccade (middle right), and feedback tone (rightmost) for error noncancelled stop signal trials (top row rasters; thick dotted line) and latency-matched no stop signal trials (middle row rasters; thin solid line) as well canceled stop signal trials (bottom row rasters; thick solid line). b, Raster and spike density plots for the representative Loss Neuron aligned on target presentation (leftmost), stop signal (middle left), saccade (middle right), and feedback tone (rightmost) for error noncancelled stop signal trials (top row rasters; thick dotted line) and latency-matched no stop signal trials (middle row rasters; thin solid line) trials as well canceled stop signal trials (bottom row rasters; thick solid line). c, Venn diagram showing the number of neurons with performance monitoring signals in the entire sample (bolded text) and only from sessions with verified perpendicular penetrations (in brackets). The overlapping regions indicate multiplexing of signals during different intervals of the trial. Equivalent minorities of Loss and Gain neurons exhibited error-related modulation. A small minority of Loss and Gain neurons modulated otherwise across intervals. A negligible fraction of neurons exhibited modulation patterns consistent with each type. These results demonstrate the existence of discrete neural types in SEF, consistent with the localization of different neuron types in different layers that have different inputs and outputs. d, Cumulative distribution of spike width for Error, Loss, and Gain neurons across all recording sessions (solid) and sessions with perpendicular penetrations (dotted). Narrow-spiking neurons (spike width < 250 µm) were commonly Loss neurons (n = 43, purple) and less commonly Error (n = 9, cyan) and Gain (n = 6, green) neurons. Narrow-spiking Loss neurons were significantly more common than narrow-spiking Gain neurons (Chi-square test, χ2(1, N = 280) = 11.11, p = 8.6 x 10−4) but not than Error neurons (χ2 (1, N = 250) = 1.79, p = 0.18).
Before we performed the layer-specific analysis, we tested the relationship between RT adaptation (RTn+1 – RTn) and the spiking of all Gain neurons (L2/3 and L5/6 combined) and of all Loss neurons (L2/3 and L5/6 combined). Neither analysis resulted in significant relationships in correct-only, error-only, or combined correct + error trials. However, when we considered the balance of activation of Gain and Loss neurons (L2/3 and L5/6 combined) with changes in RT, a significant effect was observed as reported in the main text. Next, we analyzed separately neurons in L2/3 and L5/6. However, recall that in L2/3 Gain neurons are rare, so when considering Gain neurons alone, the spike counts in L2/3 are harder to interpret. Therefore, we combined the activity of Gain and Loss neurons, appreciating that the activity of Loss neurons is dominating in L2/3. Given the antagonistic functional relationship between Gain and Loss neurons and the push-pull mechanism in the basal ganglia, we conjecture that this combination can be physiologically meaningful. Hence, we have also looked at the layer-specific relationship of RT adaptation with Gain and Loss neuron spiking. We did not find a significant influence on RT adaptation for either neuron type when considered in isolation (partial correlations to account for correlated activity across L2/3 and L5/6 (Gain neurons – L2/3: rs = −0.16, p = 0.069; L5/6: rs = −0.11, p = 0.18; Loss neurons – L2/3: rs = 0.0787, p = 0.23, L5/6: rs = −0.0213, p = 0.74). a, Zero-order Spearman’s rank correlation plots of RT adaptation as a function of the activity of Gain and Loss neurons sampled in L2/3 (top) and L5/6 (bottom), complementing Figure 6. b, Zero-order Spearman’s rank correlation plots of RT adaptation as a function of the activity of Error neurons sampled in L2/3 (top) and L5/6 (bottom), complementing Figure 3. In both panels, each point plots the average RT adaptation and associated spike count in one of 20 bins, ranked by magnitude of RT adaptation, with equal numbers of trials per bin. Thus, the number of points plotted equals 20 per session for the sessions with non-zero spike counts in both L2/3 and L5/6 (panel a: 13 sessions, panel b: 6 sessions). Normalized RT adaptation was correlated with the normalized difference in Loss and Gain neuron spiking (Aloss - Again) and with Error neuron spiking in both L2/3 and L5/6. The associated rS and p values are written in black above each panel. Statistically significant relationships are highlighted by lines of best fit. After accounting for (1) the correlation of spike rates across L2/3 and L5/6 and (2) correct versus error trial outcome, only the variation of spike rate in L2/3 predicted RT adaptation. The associated rS and p values are written in red above each panel. For neither condition alone was the variation of RT adaptation across trials correlated with the variation of spiking activity (Partial-rank correlation, controlling for correlated variability of ALoss – AGain across L2/3 and L5/6: Correct trials – L2/3: rs(257) = 0.11, p = 0.10; L5/6: rs(257) = 0.09, p = 0.18; and for Error trials – L2/3: rs(257) = −0.03, p = 0.67; L5/6: rs(257) = 0.06, p = 0.34, and controlling for correlated variability of upper and lower layer Aerror: Correct trials – L2/3: rs(117) = 0.17, p = 0.07; L5/6: rs(117) = 0.09, p = 0.36; and for Error trials – L2/3: rs(117) = 0.15, p = 0.10; L5/6: rs(117) = −0.03, p = 0.73).
About this article
Cite this article
Sajad, A., Godlove, D.C. & Schall, J.D. Cortical microcircuitry of performance monitoring. Nat Neurosci 22, 265–274 (2019). https://doi.org/10.1038/s41593-018-0309-8
This article is cited by
The unknown but knowable relationship between Presaccadic Accumulation of activity and Saccade initiation
Journal of Computational Neuroscience (2021)
Nature Communications (2020)