Figure 1.
Illustration of theoretical models to explain the pattern of covariation between tasks.
Across all figures, experimental tasks are presented in the top row, in grey. Factors are shown in the lower row, in colour. The “Intelligence” factor (purple) includes the effect of RPM, FSIQ or both, depending on the variable and group. These models describe the significant contribution of a given factor (i.e., intelligence or other factor) to the variance of any given perceptual task performance. A. Generic model. Arrows from the same factor (here, intelligence) pointing towards two tasks (here, 1 and 2) indicate that the correlation between these two tasks can be explained by their common relationship with the factor, represented here as intelligence. In the example presented, the intelligence factor does not fully explain the variance of tasks 1 and 2, and a residual covariation attributed to “another factor” (orange), not dependent on intelligence, explains this residual correlation. B. (TD controls) and C. (Autistic individuals). Models that fit the observed patterns of covariation in this study for each group separately (statistics available in Table 2 and 4). The factors not dependent on intelligence, that contribute to residual covariations include: the “Unimodal Auditory Aptitude” factor (green), the “Unimodal Visual Aptitude” factor (blue) and the “Plurimodal Perceptual Aptitude” factor (orange). The “Unimodal Auditory Aptitude” factor is a common factor found in both autistic individuals and in the general TD population and explains the relationship between levels of processing within a single perceptual modality. The “Unimodal Visual Aptitude” factor is an analogue to the “Unimodal Auditory Aptitude” factor, but within the visual modality. This factor reaches significance only in the autistics group in the current study. The “Plurimodal Perceptual Aptitude” factor is different from the unimodal aptitude factors and is present only in autistic individuals. This factor is the main finding of the current study and is given the abbreviated “p-factor” label in the discussion. Full Lines: p<0.05; Dotted lines: p<0.1.
Figure 2.
Schematic representation of the study's factorial design and presentation of experimental stimuli and tasks.
The four experimental tasks are presented in each quadrant. Each task is characterized by a sensory modality (visual or auditory) and by a level of cortical processing engaged during task completion (low- or mid-level). A. Luminance-contrast (LC) discrimination: gratings were presented for 753 ms each and separated by an inter-stimulus interval of 271 ms, during which a noise mask was presented to minimize spatial after effects. B. Pitch discrimination: pure tones were presented for 200 ms each and separated by an inter-stimulus interval of 212 ms. C. Block design completion: examples of minimum and maximum perceptual cohesiveness (PC) models. D. Melody discrimination: examples of a standard melody compared to contour modified and contour preserved conditions. Red arrows represent contour direction. Lines represent relationships of interest in the current study. Full lines: unimodal relationships, between levels of processing; Dotted lines: plurimodal relationships, within levels of processing.
Table 1.
Descriptive statistics for all participants including age and Wechsler's Intelligence Scale IQ (FSIQ, VIQ, NVIQ) and Raven Progressive Matrices (RPM) scores: mean (standard deviation); range.
Table 2.
Model 1 (Effect of intelligence on performance and between group differences in performances) main results: a. Wechsler's Full Scale IQ (FSIQ) or b. Raven Progressive Matrices (RPM).
Table 3.
Expected mean performance according to Model 1 at average intelligence and one SD above average intelligence a. Wechsler's Full Scale IQ (FSIQ) or b. Raven Progressive Matrices (RPM).
Table 4.
Model 2 (Between group differences in residual covariation) main results: a. Wechsler's Full Scale IQ (FSIQ), or b. Raven Progressive Matrices (RPM).