Improved assessment of sensorimotor gating in animal models relevant to ASD: A data modelling approach to quantify PrePulse Inhibition of the acoustic startle reflex

- Habituation/exhaustion occurred in all animals during PPI testing.
- Classic ANOVA analysis may lead to false positive effects.
- Mixed models and the use of a new calculation of PPI show that females Exhibit 30% less PPI compared to males.
- Animals exposed to VPA or EDCs do not show any deficit in PPI.
- Stressful test conditions and inaccurate data modeling are limiting factors for extrapolating animal PPI data to humans.

BackgroundThe PrePulse Inhibition (PPI) of the acoustic startle reflex is a neurobehavioral test frequently used in neurodevelopmental studies. Most PPI studies have used rodent models of schizophrenia; however, the currently used data analysis method does not take into account the variability present in autistic preclinical models.New methodWe propose a new data modelling approach for PPI data obtained from animals exposed to valproic acid or endocrine disruptors, using mixed modeling; and a new calculation of inhibition of the acoustic startle, which takes into account the habituation phenomenon.ResultsHabituation, or possibly exhaustion, occurred in all groups. The classic method of calculation of inhibition analysed with ANOVA indicated no group or sex effect for the overall inhibition of startle. In contrast, when analysed using mixed models, group and sex effects were observed. In addition, using the new method of calculation, both statistical analyses showed a sex effect, with females having decreased global inhibition but no group effect. ANOVA generated more false positive results for PPI in relation to prepulse intensities.Comparison with existing methodThe current classic method of analysis of PPI test is a calculation of inhibition based on average startle amplitude throughout the test session and a statistical ANOVA analysis. This method does not take into account habituation/exhaustion and within-subject and −group variability.ConclusionsThe results of this study demonstrate that use of ANOVA analysis leads to misinterpretation of PPI data in autistic preclinical models and we propose a new data analysis adapted to these models.