Adaptive behavior is guided by integrated representations of controlled and non-controlled information
- Department of Psychological and Brain Sciences, University of Iowa, United States
- Cognitive Control Collaborative, University of Iowa, United States
- Princeton Neuroscience Institute, Princeton University, United States
- Centre for Neuroscience Studies, Queen’s University, Canada
- Iowa Neuroscience Institute, University of Iowa, United States
Figures
Figure 1 with 1 supplement
Experimental design and behavioral results (N=40).
(a) Trial structure. (b) Procedure of the main experiment. The number following each stimulus indicates its trial count within a mini-block. (c) Group mean RT as a function of the ISPC effect. (d) Group mean error rate as a function of the ISPC effect. Error bars show standard errors of the mean (SEM). MI: mostly incongruent trials; MC: mostly congruent trials; incon: incongruent trials; con: congruent trials. *p<0.05; ** p<0.01; *** p<0.001.
Figure 1—figure supplement 1
ISPC effect in the first and second halves of Phase 2.
(a) Group mean error rate as a function of the ISPC effect. Error bars show standard errors of the mean (SEM). MI: mostly incongruent trials; MC: mostly congruent trials; incon: incongruent trials; con: congruent trials. (a) Group mean error rate as a function of the ISPC effect. Error bars show standard errors of the mean (SEM). MI: mostly incongruent trials; MC: mostly congruent trials; incon: incongruent trials; con: congruent trials.
Figure 2
Decoding performance including both SC and SR latent subspaces (N=40).
(a) Group average decoding accuracy of all 16 experimental conditions as a function of time after stimulus onset. Shaded regions represent SEM. The solid line above the time axis denotes the time points showing above-chance (0.0625, or 1/16) decoding accuracy (cluster-based permutation test, cluster-forming threshold p<0.001, cluster-level p<0.05). (b, c) MDS of EEG data across all experimental conditions. The label of each dot encodes the condition in the experimental design in the order of ISPC, congruency, color, and word. For example, ‘mccbr’ means the condition with MC, congruent trial, blue color, and the word ‘red’. The color of each dot denotes the ink color of each condition.
Figure 3
Decoding performance including both SC and SR latent subspaces using response-locked analysis (N=40).
(a) Group average decoding accuracy of all 16 experimental conditions as a function of time after response onset. Shaded regions represent SEM. The solid line above the time axis denotes the time points showing above-chance (0.0625, or 1/16) decoding accuracy (cluster-based permutation test, cluster-forming threshold p<0.001, cluster-level p<0.05). (b, c) MDS of EEG data across all experimental conditions. The label of each dot encodes the condition in the experimental design in the order of ISPC, congruency, color, and word. For example, ‘mccbr’ means the condition with MC, congruent trial, blue color, and the word ‘red‘. The color of each dot denotes the ink color of each condition.
Figure 4 with 4 supplements
Partially overlapping SC and SR subspaces.
(a) Simulation results (N=40) of decoding accuracy as a function of the degree of subspace overlap, subspace type, and decoder. Both SC and SR subspaces are eight-dimensional (SC: 4 colors ×MC/MI; SR: 4 words ×2 possible responses per word). The number of shared dimensions indicates how many dimensions overlap between SC and SR subspaces. Each condition label is encoded in the format of ‘subspace | decoder’. For example, ‘SC | SR’ means an SR decoder trained on the SC subspace. (b) Group average decoding accuracy over time as a function of which subspace the decoders are trained on. Shaded regions represent the SEM. Blue (orange) points denote the time points showing significantly better decoding accuracy when using the same subspace than when using the other subspace (cluster-based permutation test, cluster-forming threshold p<0.001, cluster-level p<0.05). (c) MDS of the SC subspace. Each dot represents the center of an SC class. Dot color and label encode the ink color and cognitive control state, respectively. (d) MDS of the SR subspace. Each dot represents the center of an SR class. The label and dot color encode the word meaning and associated response, respectively.
Figure 4—figure supplement 1
Illustration of similarity matrices of RSA on SR decoder trained on SR subspace.
To test whether word can be linearly separated in SR subspace, we performed the RSA with word on the SR decoder trained on the SR subspace. Note that there are two pairs of words that are not linearly separable in Figure 4d: red-blue and yellow-green. Thus, we specifically tested the separability within the two pairs using one predictor for each pair. The label of each row/column represents the condition including ISPC, color, and word. For example, ’mc.rb.r’ means the condition with MC, color red & blue, and word red. For each cell in a matrix, the color indicates whether the row and column conditions share the same factor (yellow = yes, purple = no) encoded by the matrix. For example, the cell at the seventh row and the first column in the 'word_red/blue’ matrix encodes that the seventh condition (i.e., mi.rb.r) and the first condition (i.e., mc.rb.r) share the same word association.
Figure 4—figure supplement 2
Linear separation of the word feature on SR subspace.
Group average t values of representational strength for word red & blue and word yellow & green on SR decoder trained on SR subspace. Squares below the lines indicate statistically significant time points (p<0.001, Bonferroni corrected). The results showed that within both word pairs (blue-red, yellow-green) individual words were represented above chance level. Considering that the decoders were linear, this finding indicated linear separability of the word pairs in the original SR subspace.
Figure 4—figure supplement 3
Illustration of similarity matrix of RSA.
(a) Similarity matrices of SC decoder trained on SR subspace. The label of each row/column represents the condition including ISPC, color, and word. For example, ’mc.r.rb’ means the condition with MC, color red, and word red & blue. For each cell in a matrix, the color indicates whether the row and column conditions share the same factor (yellow = yes, purple = no) encoded by the matrix. For example, the cell at the seventh row and the first column in the 'Color‘ matrix encodes that the seventh condition (i.e., mi.r.rb) and the first condition (i.e., mc.r.rb) share the same color association. (b) Similarity matrices of SR decoder trained on SC subspace. The label of each row/column represents the condition including ISPC, color, and word. For example, ‘mc.rb.r’ means the condition with MC, color red & blue, and word red.
Figure 4—figure supplement 4
Shared dimensions between SC and SR subspaces.
(a) Group average t values of representational strength for color, group, and ISPC over time on SC decoder trained on SR subspace (SR | SC) with stimulus-locked analysis. Squares below the lines indicate the significant time points (p<0.001, Bonferroni corrected). (b) Group average t values of representational strength for color, group, and ISPC over time on SR decoder trained on SC subspace (SC | SR) with stimulus-locked analysis. Squares below the lines indicate the significant time points (p<0.001, Bonferroni corrected). (c) Group average t values of representational strength for color, group, and ISPC over time on SC decoder trained on SR subspace (SR | SC) with response-locked analysis. Squares below the lines indicate the significant time points (p<0.001, Bonferroni corrected). (d) Group average t values of representational strength for color, group, and ISPC over time on SR decoder trained on SC subspace (SC | SR) with response-locked analysis. Squares below the lines indicate the significant time points (p<0.001, Bonferroni corrected).
Figure 5
Partially overlapping SC and SR subspaces using response-locked analysis.
(a) Group average decoding accuracy over time as a function of which subspace the decoders are trained on. Shaded regions represent the SEM. Blue (orange) points denote the time points showing significantly better decoding accuracy when using the same subspace than when using the other subspace (cluster-based permutation test, cluster-forming threshold p<0.001, cluster-level p<0.05). (b) MDS of the SC subspace. Each dot represents the center of an SC class. Dot color and label encode the ink color and cognitive control state, respectively. (c) MDS of the SR. Each dot represents the center of an SR class. The label and dot color encode the word meaning and associated response, respectively.
Figure 6
Split-half decoding and distance regression support separable SC and SR subspace.
(a, b) Group average decoding accuracy over time is plotted as a function of which subspace the decoders are trained on (stimulus-locked data [a], response-locked data [b]). Shaded regions represent the SEM. Blue (orange) points denote the time points showing significantly better decoding accuracy when using the same subspace than when using the other subspace (cluster-based permutation test, cluster-forming threshold p<0.001, cluster-level p<0.05). (c) Slope of how much decoding accuracy changes as a function of the distance between each test trial and its closest training trial (group mean and SEM), plotted as a function of time relative to stimulus onset.
Figure 7 with 1 supplement
Simultaneous EEG representations of SC and SR associations.
(a) Group average t values of representational strength for each factor over time. Squares below the lines indicate the significant time points (cluster-based permutation test, cluster-forming threshold p<0.001, cluster-level p<0.05). (b) SC and SR association results from (a). Shaded areas denote SEM. (c) Cross-trial correlation coefficient of representational strength between SC and SR associations, plotted as a function of time after stimulus onset. Shaded areas indicate SEM. Squares below the lines indicate the time points significantly above the baseline correlation strength of shuffled data (cluster-based permutation test, cluster-forming threshold p<0.001, cluster-level p<0.05).
Figure 7—figure supplement 1
Simultaneous EEG representations of SC and SR associations.
(a) Group average t values of representational strength for each factor over time. Squares below the lines indicate the significant time points (p<0.05, Bonferroni corrected). (b) SC and SR association results from (a). Shaded areas denote SEM.
Figure 8
Simultaneous EEG representations of SC and SR associations using response-locked analysis.
(a) Group average t values of representational strength for each factor over time. Squares below the lines indicate the significant time points (cluster-based permutation test, cluster-forming threshold p<0.001, cluster-level p<0.05). (b) SC and SR association results from (a). Shaded areas denote SEM. (c) Cross-trial correlation coefficient of representational strength between SC and SR associations, plotted as a function of time after response onset. Shaded areas indicate SEM. Squares below the lines indicate the time points significantly above the baseline correlation strength of shuffled data (cluster-based permutation test, cluster-forming threshold p<0.001, cluster-level p<0.05). The dashed line shows baseline correlation using shuffled data.
Figure 9 with 1 supplement
Both the strengths of SC and SR associations are correlated with RT.
(a) Group average t-values for each factor predicting RT in the LMM analysis. (b) SC and SR association results from (a).
Figure 9—figure supplement 1
The predictions of SC and SR effects on the RT on frequent and infrequent trials.
Group average t values of each factor predicting RT in the LMM analysis.
Figure 10 with 1 supplement
Correlations of SC and SR association strengths with RT using response-locked analysis.
(a) Group average t-values for each factor predicting RT in the LMM analysis. (b) SC and SR association results from (a).
Figure 10—figure supplement 1
The predictions of SC and SR effects on the RT on frequent and infrequent trials using response-locked analysis.
Group average t values of each factor predicting RT in the LMM analysis.
Figure 11
Illustration of linear discriminatory analysis (LDA).
Figure 12
Illustration of RSA.
The label of each row/column represents the condition in the experimental design including ISPC, congruency, color, and word. For example, ‘mccbb’ means the condition with MC, congruent trial, color blue, and word blue. For each cell in a matrix, the color indicates whether the row and column conditions share the same factor (yellow = yes, blue = no) encoded by the matrix. For example, the cell at the fourth row and the second column in the ‘SC’ matrix encodes that the fourth condition (i.e., mcibr) and the second condition (i.e., mccbb) share the same SC association.
Author response image 1
Shuffling analyses with stimulus-locked data support separable SC and SR subspace.
(a) Group average decoding accuracy of all 16 experimental conditions as a function of time after stimulus onset. Squares below the lines indicate the significant time points between real data and shuffled data (cluster-based permutation test, cluster-forming threshold p < 0.001, cluster-level p < 0.05). (b) Group average t values of representational strength for each factor over time. Squares below the lines indicate the significant time points between real data and shuffled data (cluster-based permutation test, cluster-forming threshold p < 0.001, cluster-level p < 0.05). (c) SC and SR association results from Fig. 1b.
Author response image 2
Shuffling analyses with response-locked data support separable SC and SR subspace.
(a) Group average decoding accuracy of all 16 experimental conditions as a function of time after stimulus onset. Squares below the lines indicate the significant time points between real data and shuffled data (cluster-based permutation test, cluster-forming threshold p < 0.001, cluster-level p < 0.05). (b) Group average t values of representational strength for each factor over time. Squares below the lines indicate the significant time points between real data and shuffled data (cluster-based permutation test, cluster-forming threshold p < 0.001, cluster-level p < 0.05). (c) SC and SR association results from Fig. 2b.
Tables
Appendix 1—table 1
Descriptive statistics of behavioral results.
Data are reported in the format of group mean (SEM). MI: mostly incongruent trials; MC: mostly congruent trials; incon: incongruent trials; con: congruent trials.
| Phase 1 | Phase 2 | Phase 3 | |||||
|---|---|---|---|---|---|---|---|
| MC | MI | MC | MI | MC | MI | ||
| RTs (ms) | con | 640.38 (10.98) | 730.58 (18.88) | 680.15 (12.20) | 651.75 (12.10) | 639.45 (10.96) | 693.87 (13.30) |
| incon | 748.65 (17.84) | 754.98 (16.49) | 745.86 (15.59) | 698.17 (14.87) | 708.65 (14.54) | 714.27 (14.62) | |
| Error rate | con | 0.02 (0.005) | 0.08 (0.010) | 0.06 (0.005) | 0.04 (0.005) | 0.04 (0.004) | 0.07 (0.008) |
| incon | 0.09 (0.010) | 0.08 (0.008) | 0.07 (0.007) | 0.05 (0.005) | 0.06 (0.008) | 0.06 (0.006) | |
Additional files
Download links
A two-part list of links to download the article, or parts of the article, in various formats.
Downloads (link to download the article as PDF)
Open citations (links to open the citations from this article in various online reference manager services)
Cite this article (links to download the citations from this article in formats compatible with various reference manager tools)
Adaptive behavior is guided by integrated representations of controlled and non-controlled information
eLife 14:RP108673.
https://doi.org/10.7554/eLife.108673.5
