Causal Dynamics Of Patient/Clinician Facial Expression Transfer Are Associated With Insula Cortex Brain-To-Brain Concordance

While patient/clinician relationships significantly impact pain outcomes, the behavioral and neural mechanisms supporting this phenomenon are unknown. Specifically, the precise role of facial expressions in contributing to analgesia and brain-to-brain communication has not been investigated. In this Hyperscan fMRI study, we explored how pain-related, directional facial communication between patient and clinician engaging in interactive pain treatment is associated with brain concordance. Dyads comprising 20 patients with fibromyalgia and 19 acupuncturists completed two synchronous fMRI scans (two different partners per participant). During fMRI, dyad members saw each other's face through closed-circuit video while interacting. The fMRI scan included a series of moderately painful and minimally/non-painful leg cuff pressure stimuli to the patient while the clinician observed. A separate fMRI scan included clinician-applied and remotely-activated electroacupuncture treatment to the patient while they received moderate cuff pressure pain. In-scanner videos were used for automated individual facial muscle unit (FMU) timeseries extraction. FMU data from pain/observation scans was fed into an interpretable machine learning classifier to determine which FMU's dynamics was most associated with pain. FMUs from the pain/treatment scans were used in conjunction with neural-network causality estimation to investigate directional non-verbal communication and their association with brain-to-brain dynamic fMRI concordance (as reported in Ellingsen et al., 2020). Causality analysis indicated a strong asymmetry in which the patients’ facial expressions elicited a range of facial expressions in clinicians, but not vice-versa. Patient lip movement FMUs were most associated with pain. Interestingly, patient-to-clinician causal influence of these FMUs on similar clinicians’ FMUs was associated with dynamic insula cortex brain-to-brain concordance, demonstrating the key role of this social mirroring and salience processing brain region with respect to facial expression mirroring between patients and clinicians. Supported by the National Institutes of Health, National Center for Complementary and Integrative Health (R61-AT009306), Norwegian Research Council /Marie Sklodowska-Curie Actions (FRICON/COFUND-240553/F20), Neuroimaging Pilot Funding Initiative at the Martinos Center for Biomedical imaging, MGH (R90DA023427), Korea Institute of Oriental Medicine (KIOM), Foundation for the Science of the Therapeutic Encounter, National Center for Research Resources (P41RR14075; CRC 1 UL1 RR025758), Harvard Clinical and Translational Science Center, Martinos Computing facilities, and National Institutes of Health (Grant Nos. S10RR023401, S10RR019307, S10RR019254, and S10RR023043). While patient/clinician relationships significantly impact pain outcomes, the behavioral and neural mechanisms supporting this phenomenon are unknown. Specifically, the precise role of facial expressions in contributing to analgesia and brain-to-brain communication has not been investigated. In this Hyperscan fMRI study, we explored how pain-related, directional facial communication between patient and clinician engaging in interactive pain treatment is associated with brain concordance. Dyads comprising 20 patients with fibromyalgia and 19 acupuncturists completed two synchronous fMRI scans (two different partners per participant). During fMRI, dyad members saw each other's face through closed-circuit video while interacting. The fMRI scan included a series of moderately painful and minimally/non-painful leg cuff pressure stimuli to the patient while the clinician observed. A separate fMRI scan included clinician-applied and remotely-activated electroacupuncture treatment to the patient while they received moderate cuff pressure pain. In-scanner videos were used for automated individual facial muscle unit (FMU) timeseries extraction. FMU data from pain/observation scans was fed into an interpretable machine learning classifier to determine which FMU's dynamics was most associated with pain. FMUs from the pain/treatment scans were used in conjunction with neural-network causality estimation to investigate directional non-verbal communication and their association with brain-to-brain dynamic fMRI concordance (as reported in Ellingsen et al., 2020). Causality analysis indicated a strong asymmetry in which the patients’ facial expressions elicited a range of facial expressions in clinicians, but not vice-versa. Patient lip movement FMUs were most associated with pain. Interestingly, patient-to-clinician causal influence of these FMUs on similar clinicians’ FMUs was associated with dynamic insula cortex brain-to-brain concordance, demonstrating the key role of this social mirroring and salience processing brain region with respect to facial expression mirroring between patients and clinicians. Supported by the National Institutes of Health, National Center for Complementary and Integrative Health (R61-AT009306), Norwegian Research Council /Marie Sklodowska-Curie Actions (FRICON/COFUND-240553/F20), Neuroimaging Pilot Funding Initiative at the Martinos Center for Biomedical imaging, MGH (R90DA023427), Korea Institute of Oriental Medicine (KIOM), Foundation for the Science of the Therapeutic Encounter, National Center for Research Resources (P41RR14075; CRC 1 UL1 RR025758), Harvard Clinical and Translational Science Center, Martinos Computing facilities, and National Institutes of Health (Grant Nos. S10RR023401, S10RR019307, S10RR019254, and S10RR023043).