Limits...
A Sensitive and Specific Neural Signature for Picture-Induced Negative Affect.

Chang LJ, Gianaros PJ, Manuck SB, Krishnan A, Wager TD - PLoS Biol. (2015)

Bottom Line: The signature was comprised of mesoscale patterns spanning multiple cortical and subcortical systems, with no single system necessary or sufficient for predicting experience.Furthermore, it was not reducible to activity in traditional "emotion-related" regions (e.g., amygdala, insula) or resting-state networks (e.g., "salience," "default mode").Overall, this work identifies differentiable neural components of negative emotion and pain, providing a basis for new, brain-based taxonomies of affective processes.

View Article: PubMed Central - PubMed

Affiliation: Department of Psychology & Neuroscience, University of Colorado, Boulder, Colorado, United States of America.

ABSTRACT
Neuroimaging has identified many correlates of emotion but has not yet yielded brain representations predictive of the intensity of emotional experiences in individuals. We used machine learning to identify a sensitive and specific signature of emotional responses to aversive images. This signature predicted the intensity of negative emotion in individual participants in cross validation (n =121) and test (n = 61) samples (high-low emotion = 93.5% accuracy). It was unresponsive to physical pain (emotion-pain = 92% discriminative accuracy), demonstrating that it is not a representation of generalized arousal or salience. The signature was comprised of mesoscale patterns spanning multiple cortical and subcortical systems, with no single system necessary or sufficient for predicting experience. Furthermore, it was not reducible to activity in traditional "emotion-related" regions (e.g., amygdala, insula) or resting-state networks (e.g., "salience," "default mode"). Overall, this work identifies differentiable neural components of negative emotion and pain, providing a basis for new, brain-based taxonomies of affective processes.

No MeSH data available.


Experimental paradigm and analysis overview.Panel A depicts the sequence of events for a given trial. Participants view an initial fixation cross and then are instructed to look at the picture (compared to reappraise). Participants then see a photo and are asked to rate how negative they feel on a likert scale of 1–5. Panel B illustrates the temporal data reduction for each rating level using voxel-wise univariate analysis and an assumed hemodynamic response function. Panel C: these voxels are then treated as features and trained to predict ratings using LASSO-PCR with leave-one-subject-out cross validation. Subject’s data for each rating is concatenated across participants. Panel D: this multivoxel weight map pattern can be tested on new data using matrix multiplication to produce a scalar affective rating prediction. Panel E: we calculated two different types of classification accuracy: (a) the ability to discriminate between high (rating = 5) and low (rating = 1) affective ratings and (b) the ability to discriminate between high affective and high pain data.
© Copyright Policy
Related In: Results  -  Collection

License
getmorefigures.php?uid=PMC4476709&req=5

pbio.1002180.g001: Experimental paradigm and analysis overview.Panel A depicts the sequence of events for a given trial. Participants view an initial fixation cross and then are instructed to look at the picture (compared to reappraise). Participants then see a photo and are asked to rate how negative they feel on a likert scale of 1–5. Panel B illustrates the temporal data reduction for each rating level using voxel-wise univariate analysis and an assumed hemodynamic response function. Panel C: these voxels are then treated as features and trained to predict ratings using LASSO-PCR with leave-one-subject-out cross validation. Subject’s data for each rating is concatenated across participants. Panel D: this multivoxel weight map pattern can be tested on new data using matrix multiplication to produce a scalar affective rating prediction. Panel E: we calculated two different types of classification accuracy: (a) the ability to discriminate between high (rating = 5) and low (rating = 1) affective ratings and (b) the ability to discriminate between high affective and high pain data.

Mentions: We used Least Absolute Shrinkage and Selection Operator and Principle Components Regression (LASSO-PCR) [35,53] to identify a distributed Picture Induced Negative Emotion Signature (PINES) that monotonically increased with increasing affective ratings in leave-one-subject-out cross validated analyses (n = 121). To apply the model to data from individual test subjects in both cross validation (n = 121) and separate hold-out test datasets (n = 61), we calculated the pattern response—the dot product of the PINES weight map and the test image—for individual subjects’ activation maps for each of 5 levels of reported negative emotion (see Fig 1). The resulting continuous values reflect the predicted intensity of negative emotion for a given activation map. We used these values to classify which of two conditions elicited a stronger negative emotion for an individual (a “forced-choice” test) [35], providing accuracy estimates (Fig 1E). We also used similar classification tests, described below, to evaluate the sensitivity and specificity of PINES responses to negative emotion versus pain. We focus primarily on results for the test sample, as it was completely independent of all model-training procedures and provides the strongest evidence for generalizability [54].


A Sensitive and Specific Neural Signature for Picture-Induced Negative Affect.

Chang LJ, Gianaros PJ, Manuck SB, Krishnan A, Wager TD - PLoS Biol. (2015)

Experimental paradigm and analysis overview.Panel A depicts the sequence of events for a given trial. Participants view an initial fixation cross and then are instructed to look at the picture (compared to reappraise). Participants then see a photo and are asked to rate how negative they feel on a likert scale of 1–5. Panel B illustrates the temporal data reduction for each rating level using voxel-wise univariate analysis and an assumed hemodynamic response function. Panel C: these voxels are then treated as features and trained to predict ratings using LASSO-PCR with leave-one-subject-out cross validation. Subject’s data for each rating is concatenated across participants. Panel D: this multivoxel weight map pattern can be tested on new data using matrix multiplication to produce a scalar affective rating prediction. Panel E: we calculated two different types of classification accuracy: (a) the ability to discriminate between high (rating = 5) and low (rating = 1) affective ratings and (b) the ability to discriminate between high affective and high pain data.
© Copyright Policy
Related In: Results  -  Collection

License
Show All Figures
getmorefigures.php?uid=PMC4476709&req=5

pbio.1002180.g001: Experimental paradigm and analysis overview.Panel A depicts the sequence of events for a given trial. Participants view an initial fixation cross and then are instructed to look at the picture (compared to reappraise). Participants then see a photo and are asked to rate how negative they feel on a likert scale of 1–5. Panel B illustrates the temporal data reduction for each rating level using voxel-wise univariate analysis and an assumed hemodynamic response function. Panel C: these voxels are then treated as features and trained to predict ratings using LASSO-PCR with leave-one-subject-out cross validation. Subject’s data for each rating is concatenated across participants. Panel D: this multivoxel weight map pattern can be tested on new data using matrix multiplication to produce a scalar affective rating prediction. Panel E: we calculated two different types of classification accuracy: (a) the ability to discriminate between high (rating = 5) and low (rating = 1) affective ratings and (b) the ability to discriminate between high affective and high pain data.
Mentions: We used Least Absolute Shrinkage and Selection Operator and Principle Components Regression (LASSO-PCR) [35,53] to identify a distributed Picture Induced Negative Emotion Signature (PINES) that monotonically increased with increasing affective ratings in leave-one-subject-out cross validated analyses (n = 121). To apply the model to data from individual test subjects in both cross validation (n = 121) and separate hold-out test datasets (n = 61), we calculated the pattern response—the dot product of the PINES weight map and the test image—for individual subjects’ activation maps for each of 5 levels of reported negative emotion (see Fig 1). The resulting continuous values reflect the predicted intensity of negative emotion for a given activation map. We used these values to classify which of two conditions elicited a stronger negative emotion for an individual (a “forced-choice” test) [35], providing accuracy estimates (Fig 1E). We also used similar classification tests, described below, to evaluate the sensitivity and specificity of PINES responses to negative emotion versus pain. We focus primarily on results for the test sample, as it was completely independent of all model-training procedures and provides the strongest evidence for generalizability [54].

Bottom Line: The signature was comprised of mesoscale patterns spanning multiple cortical and subcortical systems, with no single system necessary or sufficient for predicting experience.Furthermore, it was not reducible to activity in traditional "emotion-related" regions (e.g., amygdala, insula) or resting-state networks (e.g., "salience," "default mode").Overall, this work identifies differentiable neural components of negative emotion and pain, providing a basis for new, brain-based taxonomies of affective processes.

View Article: PubMed Central - PubMed

Affiliation: Department of Psychology & Neuroscience, University of Colorado, Boulder, Colorado, United States of America.

ABSTRACT
Neuroimaging has identified many correlates of emotion but has not yet yielded brain representations predictive of the intensity of emotional experiences in individuals. We used machine learning to identify a sensitive and specific signature of emotional responses to aversive images. This signature predicted the intensity of negative emotion in individual participants in cross validation (n =121) and test (n = 61) samples (high-low emotion = 93.5% accuracy). It was unresponsive to physical pain (emotion-pain = 92% discriminative accuracy), demonstrating that it is not a representation of generalized arousal or salience. The signature was comprised of mesoscale patterns spanning multiple cortical and subcortical systems, with no single system necessary or sufficient for predicting experience. Furthermore, it was not reducible to activity in traditional "emotion-related" regions (e.g., amygdala, insula) or resting-state networks (e.g., "salience," "default mode"). Overall, this work identifies differentiable neural components of negative emotion and pain, providing a basis for new, brain-based taxonomies of affective processes.

No MeSH data available.