Facing Challenges in Differential Classical Conditioning Research: Benefits of a Hybrid Design for Simultaneous Electrodermal and Electroencephalographic Recording

Abstract

Several challenges make it difficult to simultaneously investigate central and autonomous nervous system correlates of conditioned stimulus (CS) processing in classical conditioning paradigms. Such challenges include, for example, the discrepant requirements of electroencephalography (EEG) and electrodermal activity (EDA) recordings with regard to multiple repetitions of conditions and sufficient trial duration. Here, we propose a MultiCS conditioning set-up, in which we increased the number of CSs, decreased the number of learning trials, and used trials of short and long durations for meeting requirements of simultaneous EEG-EDA recording in a differential aversive conditioning task. Forty-eight participants underwent MultiCS conditioning, in which four neutral faces (CS+) were paired four times each with aversive electric stimulation (unconditioned stimulus) during acquisition, while four different neutral faces (CS-) remained unpaired. When comparing after relative to before learning measurements, EEG revealed an enhanced centro-posterior positivity to CS+ vs. CS- during 368-600 ms, and subjective ratings indicated CS+ to be less pleasant and more arousing than CS-. Furthermore, changes in CS valence and arousal were strong enough to bias subjective ratings when faces of CS+/CS- identity were displayed with different emotional expression (happy, angry) in a post-experimental behavioral task. In contrast to a persistent neural and evaluative CS+/CS- differentiation that sustained multiple unreinforced CS presentations, electrodermal differentiation was rapidly extinguished. Current results suggest that MultiCS conditioning provides a promising paradigm for investigating pre-post-learning changes under minimal influences of extinction and overlearning of simple stimulus features. Our data also revealed methodological pitfalls, such as the possibility of occurring artifacts when combining different acquisition systems for central and peripheral psychophysiological measures.

Keywords: EEG; MultiCS conditioning; affective learning; emotion; skin conductance.

Figure 1

Experimental procedure and differential classical conditioning. (A) The experimental procedure consisted of three different parts: The first and third parts included evaluative (SAM-ratings) and behavioral (post-experimental contingency report) measures of classical conditioning (boxes in light gray). In the second part, differential aversive conditioning was carried out, while electroencephalography (EEG) and skin conductance changes were recorded (box in dark gray). (B) Differential aversive conditioning consisted of three continuous phases (habituation, acquisition, extinction). During habituation and extinction, all conditioned stimuli (CSs) were shown without presentation of the aversive electric stimulation (UCS). During acquisition, four of the eight neutral faces (CS+; solid frame) were paired four times each with the UCS, while the other four neutral faces (CS−; dashed frame) were also shown four times, but never paired with the UCS. During all three phases, trials were shown with either short or long inter-trial intervals (ITIs) to allow for multiple repetitions in EEG but also for sufficient autonomic recovery. Note that in accordance with NimStim publication guidelines, only one model (#28) from the NimStim database is displayed. The other image shows one of the authors, but was not used as stimulus in the present study.

Figure 2

Electrodermal reactivity during habituation and extinction. (A) SC changes showing the electrodermal CS+/CS− differentiation across tested SC measurements in pre- and post-learning phases. In solid, dark blue, the last four SC measurements in habituation phase; in dashed, light blue the first four SC measurements during extinction trials. (B) Averaged electrodermal activity for the CS− and CS+, over the SC measurements selected for the repeated measures ANOVA described in this section (i.e., the last two SC measurements during habituation trials and the first two SC measurements during extinction).

Figure 3

Electroencephalography (EEG). (A) The global power plot shows the mean activation of all sensors in response to aversively paired (CS+) faces after (red solid line) and before (red dotted line) learning as well as in response to unpaired (CS−) faces after (blue solid line) and before (blue dotted line) learning from 150 ms before to 600 ms after stimulus onset. The dark gray box marks the time window (0–70 ms) in which a trigger-related technical artifact occurred. Analysis was carried out across a typical time window of the late positive potential (LPP) ranging from 300 to 600 ms. The starting point of the time window (368–600 ms) in which the significant effect emerged is marked by a bold line. (B) Global power plot displaying the mean activation across significant sensors only, with labeling being equivalent to A. (C) t-Values for the comparison of ΔCS+ (CS+_post minus CS+_pre) and ΔCS− (CS−_post minus CS−_pre) are plotted onto standard heads shown from back view. Black circles mark the sensor group displaying the significant effect during the selected time window (368–600 ms). (D) ΔCS+ and (E) ΔCS− are plotted onto standard heads shown from back view.

Figure 4

Subjective ratings of hedonic valence and emotional arousal. Mean (A) hedonic valence and (B) emotional arousal ratings acquired pre- and post-learning are shown for the neutral faces (central plot) which were paired (CS+; diamonds and solid red line) or remained unpaired (CS−; circles and dashed blue line) with UCS (electrical shocks) during the differential aversive conditioning procedure, as well as for the faces of CS+ and CS− identity with angry (left plot) or happy (right plot) expression. Stars depict the significance of the t-test comparing mean CS+ and CS− ratings after conditioning, with a greater number of stars depicting a significance value of p < 0.05, 0.01, 0.005, and.001, respectively.