Abstract
We examined the constructs of behavioral response and threshold of Dunn’s four-quadrant model of sensory processing. We tested if these axes, as measured by the Adult and Adolescent Sensory Profile, are ordinal ranges and if variation is consistent with external measures of characteristics that may also reflect sensory processing, Eysenck's personality model, and somatosensory event-related potentials and sensory gating.
Primary Author and Speaker: Alexia Metz
Contributing Authors: Daniella Boling, Ashley DeVore, Holly Holladay, Karen Karmol, Jo Fu Liao
Dunn’s four-quadrant model of sensory processing (1997) has clinical utility, and the associated standardized assessments have strong psychometrics. In this study, we examined the constructs that underlie the model, behavioral response (BR) and threshold (T). We investigated whether these axes, as measured by the Adult/Adolescent Sensory Profile (A/ASP, Brown & Dunn 2002), are ordinal ranges and if variation along the ranges is consistent with variation in external measures of characteristics that may also reflect sensory processing, Eysenck's personality model (Sato, 2005) and somatosensory event related potentials (SERP) and sensory gating (Davies & Gavin 2007).
We recruited a convenience sample of 139 healthy adults. They completed the A/ASP and Eysenck's Personality Questionaire, Brief Version (EPQBV). We used median nerve stimulation to activate somatosensory pathways and recorded responses using EEG, measuring the average amplitude of the P50 waveform at Cz in response to 200 stimuli in a 5-Hz train and P50 gating in response to second stimuli in 200 pairs (500 ms interval).
We converted raw scores of the A/ASP to axis scores. There are 30 items each quadrant, rated on a 5-point scale. To prevent a gap, we subtracted 1 from each response. Passive BR and Low T axis scores were multiplied by -1. When totaled, axes could range from -120 to 120. There were strong alignments between the categorization of BR (Passive or Active) and T (Low or High) of the A/ASP quadrant with the highest raw score and the respective calculated axis score (p<.001, Fisher’s Exact Test for both axes, Cramer’s V =.608 behavioral response, .555 threshold). We plotted the calculated BR axis score against the calculated T axis score to derive a quadrant. The alignment between the quadrant with the highest raw score and the derived quadrants was significant (χ2(6)=72.650, p<.001, Cramer’s V=.513). We assessed whether calculated scores along each axis would demonstrate ordinal scaling. Categorizing participants by the standard deviation (SD) of their highest raw score, we found that calculated BR axis scores did not differ in means but those for threshold axis scores did, going from lowest for the +2 SD Low T to highest for the +1 SD High T (F(3,120)=4.198, p=.007).
We compared variation in BR and T with variation in introversion/extraversion (I/E) and neuroticism (N) of axes of Eysenck’s personality model, respectivley. There was no difference in I/E scores for Passive and Active BR patterns (-3.0±7.2 (n=7) vs. 1.8±8.0 (n=131), t(136)=-1.554). There was a small, significant positive correlation between calculated T axis scores and scores on the N scale of the EPQ-BV (Pearson’s R=.181, r 2 =.033, p=.034).
We compared variation in calculated T axis scores to variation in SERP and sensory gating. The P50 for participants with Low T patterns was 0.98±0.4 µV and for participants with High T patterns 0.76±1.3 µV (t(33)=-.433, p=.668). The calculated threshold and the amplitude of the P50 had a nonsignificant weak negative correlation (R=-.089, r 2 =.007, p=.611). The gating ratio for participants with Low T patterns was 0.82±0.42 and for participants with High T patterns was 0.77±.41 (t(72)=.335, p=.738). The calculated T axis scores and the P50 gating ratio had nonsignificant weak negative correlation (R=-.036, r 2 =.001, p=.764, n=74).
The findings suggest that the behavioral response axis of Dunn's model may not reflect a construct with an ordinal range, but the threshold axis may. The range of the threshold axis may correlate with Eysenck's concept of neuroticism, but not with somatosensory ERP and sensory gating, in healthy adults. Conclusions are limited by having few participants with passive behavior regulation and low threshold patterns of processing.
Dunn, W. (1997). The impact of sensory processing abilities on the daily lives of young children and their families: A conceptual model. Infant & Young Children, 9(4), 23-35.
Sato, T. (2007). The psychometric properties of the Eysenck Personaity Questionnaire-Brief Version. New Psychological Tests and Testing Research, 105-120.
Davies, P. L., & Gavin, W. J. (2007). Validating the diagnosis of sensory processing disorders using EEG technology. The American Journal of Occupational Therapy, 61(2), 176-189.