Limits...
Decoding the Traumatic Memory among Women with PTSD: Implications for Neurocircuitry Models of PTSD and Real-Time fMRI Neurofeedback.

Cisler JM, Bush K, James GA, Smitherman S, Kilts CD - PLoS ONE (2015)

Bottom Line: By contrast, a multivariate pattern analysis (MVPA) approach towards identifying the neural mechanisms engaged during trauma memory recall would entail testing whether a multivariate set of brain regions is reliably predictive of (i.e., discriminates) whether an individual is engaging in trauma or non-trauma memory recall.Cross-validation classification accuracy was significantly above chance for all methodological permutations tested; mean accuracy across participants was 76% for the methodological parameters selected as optimal for both efficiency and accuracy.Classification accuracy was significantly better for a voxel-wise approach relative to voxels within restricted regions-of-interest (ROIs); classification accuracy did not differ when using PTSD-related ROIs compared to randomly generated ROIs.

View Article: PubMed Central - PubMed

Affiliation: Brain Imaging Research Center, Department of Psychiatry, University of Arkansas for Medical Sciences, Little Rock, Arkansas, United States of America.

ABSTRACT
Posttraumatic Stress Disorder (PTSD) is characterized by intrusive recall of the traumatic memory. While numerous studies have investigated the neural processing mechanisms engaged during trauma memory recall in PTSD, these analyses have only focused on group-level contrasts that reveal little about the predictive validity of the identified brain regions. By contrast, a multivariate pattern analysis (MVPA) approach towards identifying the neural mechanisms engaged during trauma memory recall would entail testing whether a multivariate set of brain regions is reliably predictive of (i.e., discriminates) whether an individual is engaging in trauma or non-trauma memory recall. Here, we use a MVPA approach to test 1) whether trauma memory vs neutral memory recall can be predicted reliably using a multivariate set of brain regions among women with PTSD related to assaultive violence exposure (N=16), 2) the methodological parameters (e.g., spatial smoothing, number of memory recall repetitions, etc.) that optimize classification accuracy and reproducibility of the feature weight spatial maps, and 3) the correspondence between brain regions that discriminate trauma memory recall and the brain regions predicted by neurocircuitry models of PTSD. Cross-validation classification accuracy was significantly above chance for all methodological permutations tested; mean accuracy across participants was 76% for the methodological parameters selected as optimal for both efficiency and accuracy. Classification accuracy was significantly better for a voxel-wise approach relative to voxels within restricted regions-of-interest (ROIs); classification accuracy did not differ when using PTSD-related ROIs compared to randomly generated ROIs. ROI-based analyses suggested the reliable involvement of the left hippocampus in discriminating memory recall across participants and that the contribution of the left amygdala to the decision function was dependent upon PTSD symptom severity. These results have methodological implications for real-time fMRI neurofeedback of the trauma memory in PTSD and conceptual implications for neurocircuitry models of PTSD that attempt to explain core neural processing mechanisms mediating PTSD.

No MeSH data available.


Related in: MedlinePlus

a) scatter plots for a single participant showing the spatial correlation between SVM feature weights for each voxel when trained using a single run with SVM feature weights when trained using two runs (top) and correlation between SVM feature weights when trained using a single run and SVM feature weights when trained using 5 runs. b) spatial correlation across participants for the voxelwise SVM feature weights between the different models.
© Copyright Policy
Related In: Results  -  Collection

License
getmorefigures.php?uid=PMC4524593&req=5

pone.0134717.g002: a) scatter plots for a single participant showing the spatial correlation between SVM feature weights for each voxel when trained using a single run with SVM feature weights when trained using two runs (top) and correlation between SVM feature weights when trained using a single run and SVM feature weights when trained using 5 runs. b) spatial correlation across participants for the voxelwise SVM feature weights between the different models.

Mentions: We then assessed how similar (using r-to-z transformed correlation coefficients) the SVM feature weights were for a given participant across the models using different numbers of memory recall repetitions. Fig 2A depicts this approach for a single participant. This participant’s SVM feature weights across GM voxels for the model trained using the first run was highly similar to the SVM feature weights for the model trained using the first two runs (r = .65); by contrast, similarity between SVM feature weights for the model trained with the first run and the model trained with all five runs for this same participant was less robust (r = .29). When this approach is repeated across all participants (Fig 2B), SVM feature weights trained with 3–4 memory repetitions demonstrated greatest similarity across models. This was statistically tested by examining the median similarity for each model type (e.g., for the three run model, the median similarity of the three run model with all other models), which demonstrated that the median similarity across models for the model trained with the just the first run was significantly less than the similarity across models for the two run model (p < .001). The two run model had significantly less similarity across models than the three run model (p< .001), the three run model overall similarity did not differ from the four run model similarity (p = .97), and the four run model overall similarity was significantly greater than the five run model overall similarity (p < .001).


Decoding the Traumatic Memory among Women with PTSD: Implications for Neurocircuitry Models of PTSD and Real-Time fMRI Neurofeedback.

Cisler JM, Bush K, James GA, Smitherman S, Kilts CD - PLoS ONE (2015)

a) scatter plots for a single participant showing the spatial correlation between SVM feature weights for each voxel when trained using a single run with SVM feature weights when trained using two runs (top) and correlation between SVM feature weights when trained using a single run and SVM feature weights when trained using 5 runs. b) spatial correlation across participants for the voxelwise SVM feature weights between the different models.
© Copyright Policy
Related In: Results  -  Collection

License
Show All Figures
getmorefigures.php?uid=PMC4524593&req=5

pone.0134717.g002: a) scatter plots for a single participant showing the spatial correlation between SVM feature weights for each voxel when trained using a single run with SVM feature weights when trained using two runs (top) and correlation between SVM feature weights when trained using a single run and SVM feature weights when trained using 5 runs. b) spatial correlation across participants for the voxelwise SVM feature weights between the different models.
Mentions: We then assessed how similar (using r-to-z transformed correlation coefficients) the SVM feature weights were for a given participant across the models using different numbers of memory recall repetitions. Fig 2A depicts this approach for a single participant. This participant’s SVM feature weights across GM voxels for the model trained using the first run was highly similar to the SVM feature weights for the model trained using the first two runs (r = .65); by contrast, similarity between SVM feature weights for the model trained with the first run and the model trained with all five runs for this same participant was less robust (r = .29). When this approach is repeated across all participants (Fig 2B), SVM feature weights trained with 3–4 memory repetitions demonstrated greatest similarity across models. This was statistically tested by examining the median similarity for each model type (e.g., for the three run model, the median similarity of the three run model with all other models), which demonstrated that the median similarity across models for the model trained with the just the first run was significantly less than the similarity across models for the two run model (p < .001). The two run model had significantly less similarity across models than the three run model (p< .001), the three run model overall similarity did not differ from the four run model similarity (p = .97), and the four run model overall similarity was significantly greater than the five run model overall similarity (p < .001).

Bottom Line: By contrast, a multivariate pattern analysis (MVPA) approach towards identifying the neural mechanisms engaged during trauma memory recall would entail testing whether a multivariate set of brain regions is reliably predictive of (i.e., discriminates) whether an individual is engaging in trauma or non-trauma memory recall.Cross-validation classification accuracy was significantly above chance for all methodological permutations tested; mean accuracy across participants was 76% for the methodological parameters selected as optimal for both efficiency and accuracy.Classification accuracy was significantly better for a voxel-wise approach relative to voxels within restricted regions-of-interest (ROIs); classification accuracy did not differ when using PTSD-related ROIs compared to randomly generated ROIs.

View Article: PubMed Central - PubMed

Affiliation: Brain Imaging Research Center, Department of Psychiatry, University of Arkansas for Medical Sciences, Little Rock, Arkansas, United States of America.

ABSTRACT
Posttraumatic Stress Disorder (PTSD) is characterized by intrusive recall of the traumatic memory. While numerous studies have investigated the neural processing mechanisms engaged during trauma memory recall in PTSD, these analyses have only focused on group-level contrasts that reveal little about the predictive validity of the identified brain regions. By contrast, a multivariate pattern analysis (MVPA) approach towards identifying the neural mechanisms engaged during trauma memory recall would entail testing whether a multivariate set of brain regions is reliably predictive of (i.e., discriminates) whether an individual is engaging in trauma or non-trauma memory recall. Here, we use a MVPA approach to test 1) whether trauma memory vs neutral memory recall can be predicted reliably using a multivariate set of brain regions among women with PTSD related to assaultive violence exposure (N=16), 2) the methodological parameters (e.g., spatial smoothing, number of memory recall repetitions, etc.) that optimize classification accuracy and reproducibility of the feature weight spatial maps, and 3) the correspondence between brain regions that discriminate trauma memory recall and the brain regions predicted by neurocircuitry models of PTSD. Cross-validation classification accuracy was significantly above chance for all methodological permutations tested; mean accuracy across participants was 76% for the methodological parameters selected as optimal for both efficiency and accuracy. Classification accuracy was significantly better for a voxel-wise approach relative to voxels within restricted regions-of-interest (ROIs); classification accuracy did not differ when using PTSD-related ROIs compared to randomly generated ROIs. ROI-based analyses suggested the reliable involvement of the left hippocampus in discriminating memory recall across participants and that the contribution of the left amygdala to the decision function was dependent upon PTSD symptom severity. These results have methodological implications for real-time fMRI neurofeedback of the trauma memory in PTSD and conceptual implications for neurocircuitry models of PTSD that attempt to explain core neural processing mechanisms mediating PTSD.

No MeSH data available.


Related in: MedlinePlus