Network Analysis of Induced Neural Plasticity Post-Acceptance and Commitment Therapy for Chronic Pain

Chronic musculoskeletal pain is a costly and prevalent condition that affects the lives of over 50 million individuals in the United States. Chronic pain leads to functional brain changes in those suffering from the condition. Not only does the primary pain network transform as the condition changes from acute to persistent pain, a state of hyper-connectivity also exists between the default mode, frontoparietal, and salience networks. Graph theory analysis has recently been used to investigate treatment-driven brain network changes. For example, current research suggests that Acceptance and Commitment Therapy (ACT) may reduce the chronic pain associated hyper-connectivity between the default mode, frontoparietal, and salience networks, as well as within the salience network. This study extended previous work by examining the associations between the three networks above and a meta-analytically derived pain network. Results indicate decreased connectivity within the pain network (including left putamen, right insula, left insula, and right thalamus) in addition to triple network connectivity changes after the four-week Acceptance and Commitment Therapy intervention.

. The seven-step timeline of each individual 90-minute ACT session. "Homework" involved continuing practice of the skill learned in the prior session, as well as completing selfmonitoring and self-reporting activities. Figure S2. A basic timeline of the entire study, starting with pre-treatment screening and ending with post-treatment assessments. Check Table S3. List of nodes presented in Figure 2 (A, B, C). The node assignment matches the 264 Power atlas nodes or 7 pain nodes found above.  74,76,95,110,111,112,113,115,122,130 Default Mode Network 175,189,191,194,195,196,199,202 Frontoparietal Network 203,204,206,208,209,211,212,215,216 Salience Network were obtained from the NIH Toolbox and were administered via iPad. Any additional assessments were administered via paper copy. This set of assessments (Table S1) was administered to each subject both pre-and post-ACT.

Supplement 1.2. Neuropsychological Assessment Analysis
Assessment data were entered into Excel spreadsheets using Qualtrics software for statistical analysis. Analyses were conducted using SAS  v.9.4. to yield measures such as mean, standard deviation, Student's T (Change), Pr > |t|, Wilcoxon Signed Rank, and Pr >= |S| for a total score pre-ACT, total score post-ACT, and the change score between the two. Positive or negative change scores indicated satisfactory results, depending on the specific assessment in question (CESD scores decreasing meant less depression, and AAQ-II scores increasing meant greater feelings of acceptance, e.g.).

Supplement 1.3. Neuropsychological Assessment and rsfMRI Data Correlation Analysis
Additionally, the fMRI data set was merged with the neuropsychological assessment data set so that any correlations between change in neural plasticity (functional connectivity measures) and change in neuropsychological health indicators (behavioral measures) could be analyzed pre-and post-ACT. An Excel spreadsheet was compiled of each subject's pre-and post-ACT values for all significant edge connections (56 total; derived from an NBS analysis between DMN, FPN, and SN with a threshold of t >2.1 instead of t >2.5; see Table S6) as well as certain significant change scores (9 reported; AAQ-II, BPI, CESD, FFMQ, PCL-5, Neuro-QoL Depression, Neuro-QoL Fatigue, Neuro-QoL Satisfaction with Social Roles and Activities, and NIH Toolbox Pattern Comparison Processing Speed; the CPAQ was not used). These 9 tests (which previously yielded significant change results; see Table S7) were selected for a more cohesive understanding of brain region activation as it corresponds to chronic pain-specific assessments. SAS  v.9.4 was again used to derive correlational data (pairwise Pearson correlation, R, and p value) between functional connectivity changes and neuropsychological assessment changes. For the first pass, all scores with a p < 0.10 were selected. The list was then cut down to seven assessment measures and 15 functional connectivity measures. For the second pass, all scores with a p < 0.0071 were selected (based on Bonferroni equation for multiple comparison correction), in addition to a p < 0.05.

Supplement 1.4. Neuropsychological Assessment Data Change Results
Improvements were found based on the assessment data from pre-to post-ACT. Of the administered tests, approximately half were found to have significant change scores (Table S7).
Positive values for the following assessments represent: higher levels of chronic pain acceptance and action (AAQ-II, CPAQ), mindfulness (FFMQ), processing speed (NIH Toolbox PCPS), and satisfaction with social roles (Neuro-QoL Satisfaction with Social Roles & Activities).

Supplement 1.5. Neuropsychological Assessment and rsfMRI Data Correlation Results
To further the investigation, the behavioral data and the resting state fMRI data were run together to search for any correlational relationships. Through multiple passes, the initial edges and assessments were narrowed down to 15 edges and seven assessments. The final pass led to the correlations found in Table S8.
13 edges were included in significant (p < 0.05) correlations with six assessments (left column and top row in Table S8). The first assessment (AAQ) was correlated with one edge representing the functional connection between right AG and left ACC. The second assessment So, the significant changes in functional connectivity, namely within cingulate cortex and parahippocampal, precentral, and frontal gyri, were correlated with the significant changes in assessments regarding pain, social satisfaction, depression, and processing speed pre-to post-ACT. To correct for multiple comparisons, the data were assessed using a p value of 0.0071 (calculated by dividing 0.05 by the final number of assessments included, seven). Only one correlation involving the right angular gyrus and left anterior cingulate was significant using the Bonferroni method described (starred in Table S8).

Supplement 1.6. Main Findings in Neuropsychological Assessment Data
Behavioral data were collected using select domains for the NIH Cognition Battery, PROMIS and Neuro-QoL on the NIH Toolbox as well as physical paper surveys (Table S1).
Approximately half of all of the assessments changed significantly. All of the significant change scores demonstrated improved scores for the nine participants (Table S7)

Supplement 1.7. Main Findings in Correlation Analysis Data
Because of the significant findings of the behavioral assessment analysis and the rsfMRI graph theory analysis, it was critical to then investigate if the two independent findings were correlated to each other. This additional correlational analysis was conducted (ultimately) using seven behavioral assessment measures, 15 rsfMRI functional connectivity measures (edges), and the SAS  v.9.4 software. Pairwise Pearson correlations exhibited 14 significant correlations that were either strong (|0.6| -|0.8|) or very strong (> |0.8|) correlations (Table S8). Within the 14 correlations, four of the assessments (BPI, FFMQ, NQ DEP, and NQ SSR) correlated with more than one edge. This indicates important functional connections related to the improved assessment scores. Specifically, the BPI scores were significantly correlated with five connections (involving frontal gyri, cingulate, and more), demonstrating multiple important interactions between ACT-induced neural plasticity and subjective perception of pain severity.
Of importance is the connection between the right angular gyrus and the left anterior cingulate that was significantly correlated with the AAQ-II when correcting for multiple comparisons (using the Bonferroni method). At a p value of 0.0024 and R = 0.86811, this is a very strong correlation. The anterior cingulate has been previously linked to the cognitive and emotional regulation of pain processing as well as response selection, as opposed to aspects of pain intensity. In future investigations, the anterior cingulate and angular gyrus should be considered as potential underlying neural mechanisms. In addition, the behavioral measure for acceptance and action surrounding chronic pain should be studied more in depth.
Previous studies have also shown similar neurological interactions and revealed the importance of their connections. In numerous other ACT investigations with a chronic pain population, the medial frontal gyrus (found to have numerous significant correlations in the current study) has shown similar deactivation patterns. The medial frontal gyrus is namely responsible for assessing the risk of chronic pain as it manifests, possibly even before the chronicity is established [2]. When connections between the prefrontal cortex and anterior cingulate have been shown, those connections have been attributed to attentional and memory network activation as a response to the painful stimuli. In discussing the anterior cingulate, it is important to note the region's probable link to anxiety and depression (as depression measures were decreased in the current study). Additional studies have also demonstrated activation changes in the anterior cingulate after ACT interventions, further indicating its level of interaction.   Table S6. List of (56) significant edges of the triple network (t >2.1). The node assignment is listed out of the 264 Power atlas nodes. The brain regions corresponding to the nodes of each connection are listed also, with the abbreviations used in this paper.  Note. S: Wilcoxon Signed Rank; T: Student's T (Change). T scores were not reported for assessments administered by paper (AAQ-II, e.g.) because normality could not be assumed.

ROI List
Change scores represent the difference in score between the first and second timepoints (negative S and T values indicate that the score post-ACT was lower than pre-ACT, while positive S and T values indicate post-ACT scores higher than pre-ACT). Table S8. Correlations between significant assessment and edge scores.