The auditory evoked P3 and the omission evoked potential decrease with predictability: a single-trial analysis with wavelet denoising reveals individual learning curves

Jongsma, M. L. A.1, Eichele, T.2, Coenen, A. M. L.1, Hugdahl, K.2, Nordby, H.2, Van Rijn, C. M.1, and Quian Quiroga, R.3,
1NICI, Department of Biological Psychology, University of Nijmegen, the Netherlands; 2Department of Biological and Medical Psychology, Division of Cognitive Neuroscience, University of Bergen, Norway; 3Sloan-Swartz Center for Theoretical Neurobiology. California Institute of Technology, Pasadena, CA

Within the oddball paradigm, participants commonly respond to infrequently occurring targets. Thus evoked potentials (EPs) contain a large positive potential, the P3. It is well known that when target-probability increases, P3 amplitude decreases. Some have proposed that this is mainly due to a decrease in target-to-target interval. In this study we propose that the ability to predict target occurrence (which increases when probability is high) also decreases the P3.

In a first experiment targets (n=96) with a 12.5% probability, interspersed within a train of backgrounds (ISI 800 ms), were presented. Targets were either deviant stimuli (session A) or omitted stimuli (session B). Within one session, blocks of 8 consecutive targets were alternately presented at either a random or fixed position. Thus, targets presented within fixed cycles became predictable after a couple of presentations. In a second experiment the same paradigm was presented, but expanded by adding either a fixed or random distractor stimulus (probability 12.5%). Single-trial analysis by means of wavelet denoising was applied in order to investigate trial-to-trial variation.

In response to targets presented in fixed cycles, P3 amplitudes rapidly diminished (after 2-3 presentations), resulting in a steep learning curve. No such decrease was observed when targets were presented in random cycles. Preliminary results show that distractor stimuli might flatten these learning curves

By applying this “learning oddball paradigm” individual learning curves can be determined when using single-trial wavelet denoising. In addition, susceptibility towards distraction might be additionally assessed. Therefore, this paradigm could lead to a clinical useful tool.