Last updated on August 2019

Measuring Automated Behavioral Observations & Vocal Expressions (ABOVE)

Brief description of study

The investigators propose to synchronize automated measurements of behavior - head, body, and face dynamics, gaze, and vocal prosody - with simultaneous recordings of brain activity in clinically relevant contexts.

Detailed Study Description

Baylor College of Medicine (BCM), University of Pittsburgh (UPitt) and Brown University will collaborate on this research project. BCM will be the only site that will enroll study subjects and perform study assessments. UPitt and Brown will supervise data management, measurement and modeling of multimodal signals and collaborate on research design and machine learning.

Following initial screening, subjects will participate in one testing session lasting about 2 hours. The addition of these preliminary studies will allow investigators to obtain information in order to differentiate and code OCD-distress versus social anxiety and identify rituals related to OCD.

Research Material, Data, and Information to be collected:

Interviews, rating scales, video, audio, behavioral assessments, and heart rate will be obtained.

Preliminary Testing in Un-implanted Subjects: Automated Behavioral Measures of OCD-related Distress (i.e., obsessions and urge to ritualize) and Overt Rituals:

The Provocation OC task (Provoc) and the Trier Social Stress Test (TSST) will be used to evoke OC-related distress and anxiety unrelated to OCD (e.g., performance anxiety), respectively. These tasks will be preceded by a neutral baseline (the subject may be asked to describe how they traveled to the clinic to the experimenter) and a positive affect induction, the Broad-Minded Affective Coping procedure (BMAC) a clinical technique that uses recalled positive autobiographical memories and mental imagery to elicit positive affect. Because anxiety and negative affect may have carryover effects, neutral baseline and BMAC will be presented prior to the non-OCD and OCD distress inductions. In the BMAC, subjects will be instructed to recall and describe a recent positive memory/experience. The order of Provoc and TSST will be counterbalanced across the subjects. Use of a conveyor belt may be used during the Provoc tasks to present the evoking stimuli. Simultaneously recorded measures are noted in the Table attached in Section S and include: AFAR, new suite of behavioral measures (shown as shaded), hand/arm movements and heart rate.

A third provocation task will be added to elicit overt rituals (e.g., handwashing) in the clinic. It will occur following the BMAC and TSST/Provoc tasks and a relaxation break.

Subjects will be asked to wash their hands at a sink in our clinic inside an exam room where their behaviors will be recorded using hardware-synchronized cameras: one, to capture frontal view of the face for AFAR; and another, to record full body movement for CPM. The investigators are aware of only one prior study of automated coding of compulsive behavior. In that study, a computer vision approach was able to discriminate subjects with OCD from healthy subjects on the basis of handwashing videos in which regions of interest (ROIs) were defined for objects (e.g., soap dispenser) used for handwashing. The investigators will combine this ROI paradigm with a CPM approach that captures hand and body movements. Dropouts from the ritual paradigm will be replaced.

These same tasks will also be performed on up to 4 subjects that have consented to receive a DBS implant on a separate protocol. Informed consent will also be collected from these implanted subjects for the purpose of completing the tasks described above on this study.

Clinical Study Identifier: NCT03459456

Find a site near you

Start Over

Brown University

Providence, RI United States
  Connect »