PinPoint: Computer-Assisted Digital Video Coding

Background:  Clinicians treating individuals with autism are making increased use of digital video to record behavioral data. This trend is spurred in part by the increasing availability of affordable digital video cameras, technological solutions for remote monitoring via web-based cameras, and media-based information networks. Video coding refers to the process of identifying and recording events of interest within video content. Typically, a computer program displays the video, the observer views the video and records events by pressing designated keys, and the program creates a record of events and associated time stamps. Alternatives to coding the entire video include the widely used observational techniques of Partial Interval Recording (PIR) and Momentary Time Sampling (MTS). For both PIR and MTS, the video is divided into equal recording intervals (e.g., 15 seconds). With PIR the observer records whether the event occurred during each interval; with MTS, the observer records whether the event occurred during a brief (e.g., 2 s) observation period at the end of each interval. The dependent measure for each method is percent of intervals in which the event occurred.

PinPoint is a software application that can increase the efficiency of PIR and MTS for digital video. As each event is recorded, the software automatically advances the video playback to the next point at which an observation is required; the user need observe only those portions of the video that are necessary for the chosen coding technique. Technology development was supported by an STTR Technology Transfer grant from the National Institute of Child Health and Human Development.

Objectives: The objectives of this presentation are (1) to present data that directly compared accuracy and efficiency of MTS with standard procedures and with computer-assisted procedures, and (2) to offer a hands-on demonstration of the software application.

Methods:  Eighteen clinical professionals coded video samples of children with autism engaging in vocal or motor stereotypy. With the Standard procedure, the video was presented and an auditory beep at each observation period cued the observer to look at the video screen and record behavioral occurrence/non-occurrence on a prepared data sheet. With the Computer-assisted procedure, the software displayed only the observing periods and the observer pressed a Yes or No button after viewing each one.

Results:  Inter-observer agreement and coding accuracy did not differ between procedures (p > 0.50, paired t-test). Savings in mean time to complete coding sessions with the Computer-assisted procedure was 79% (p < 0.001).

Conclusions: The automatic video advance feature of the computer-assisted method did not distort the coding process in any way, and it produced a highly significant increase in efficiency. Visitors to the proposed poster/demonstration will have the opportunity to test-drive the software's coding interface and review its capacity to predict the effects of adjustments in PIR or MTS parameters.