Detecting Dmils Effects In Eda Records: A Signal-Engineering Approach

We thank our colleagues at the IGPP, Christian Scheer, Markus Binder, and Rainer Schneider, for their effective help with data preparation and transfers as well as for making data from their next DMILS study available to us for comparison. We also gratefully acknowledge IGPP funding and support of the study.

ABSTRACT: The aim of this study is to explore various means of quantitative assessment of DMILS effects as measured by electrodermal activity (EDA). The approach adopted in this study is that of signal engineering rather than that of electrophysiology: we have focused on extraction of EDA signal parameters that seem to carry the useful information on the effect as opposed to a search for an optimal assessment of the hypothetical target variable (level of arousal). Such redefinition of the task allows us to assess the effects in terms of formally well-defined, easy-to-compute signal parameters.

The signal analysis procedure was based on integration of rectified (absolute values or squared values) first or second derivatives of the EDA time series. The integral values computed over activating and calming periods are used to compute two dimensionless indices: an index of effect magnitude (Q) and an index of effect discrimination (D). The procedure was applied to two datasets originating from DMILS studies performed in Freiburg (receivers only, N = 80) and Edinburgh (receivers and agents, N= 2 x 80). The results applied to receivers' data indicate that the procedure is able to detect subtle effects with higher sensitivity for the second derivative of the EDA signals.

The assessment of the EDA DMILS effects by formal signal characteristics seems to be a viable strategy in the variety of DMILS research methods. On the other hand, the decline from a primarily physiological definition implies that the effects identified in this way could be conceptualized instead as a special case of complex PK. We consider this an advantage rather than a drawback, as further signal-oriented studies may lead to a separation of PK effects possibly acting solely on a physical level from neat DMILS effects affecting the physiological processes of the target system.

COPYRIGHT 1999 Parapsychology Press
COPYRIGHT 2004 Gale Group