Abstract
Attenuated sound–shape matching in the classic Kiki–Bouba effect in autism has already been replicated in several studies, but it remains unclear whether this attenuation is confined to low-level multisensory integration or whether it reflects a broader reduction in crossmodal priors. In this pre-registered study, we tested 68 intelligence quotient (IQ)-matched participants (31 autistic, 37 typically developing) on four established crossmodal correspondence tasks that span perceptual to conceptual mappings: (1) “Fast lemons” (object–adjective), (2) sound–size pairings with Greebles, (3) kiki–bouba variants, and (4) color–taste matches. A generalized linear mixed-effects model showed a significant main effect of autism, indicating fewer theoretically congruent choices across all tasks. Target consistency increased with age, performance intelligence quotient (PIQ) and with an artificial intelligence (AI)-derived item-difficulty index. Random slopes revealed no task-by-diagnosis interaction, suggesting that diminished correspondence strength in autism generalizes across sensory and semantic domains. These results cautiously point to domain-general accounts, such as predictive-coding “hypo-priors,” over multisensory-specific explanations, and highlight crossmodal correspondences as an intriguing window into information processing in autism.
Lay Abstract
People often make associations between information that comes from different senses – things they see, hear, and feel. For example, one of the classic findings in psychology is that most people think the made-up word “kiki” fits a spiky shape, while “bouba” matches a round one. Previous research has shown that people with autism may be less likely to make these typical associations. However, it’s not clear whether this difference is specific to that one example (sound and shape associations) or part of a wider difference in how people with autism link information from different senses. In this study, we tested 68 teenagers and adults (31 autistic and 37 non-autistic), all with similar intelligence quotient (IQ) levels, on four different tasks that involved matching images with other types of information. These included deciding whether a lemon is more “fast” or “slow”, which made-up name fits a creature of a certain size, what taste fits a color, and which name fits a shape. There were no right or wrong answers, but some choices matched what most people typically select. We found that autistic participants across different tasks were less likely than non-autistic participants to choose the most common or expected pairings. These findings suggest that autistic people may form different types of sensory connections, not just in sound–shape tasks but across a broader range of sensory and conceptual experiences. What does this tell us about autistic minds? While more research is needed to know for sure, our findings are consistent with a theory that autistic people may rely less on previous experiences when processing new information. This theory is known as the “hypo-priors” account, which suggests that autistic perception is less shaped by learned expectations and more by the raw details of what’s currently being sensed.
Get full access to this article
View all access options for this article.
References
Supplementary Material
Please find the following supplemental material available below.
For Open Access articles published under a Creative Commons License, all supplemental material carries the same license as the article it is associated with.
For non-Open Access articles published, all supplemental material carries a non-exclusive license, and permission requests for re-use of supplemental material or any part of supplemental material shall be sent directly to the copyright owner as specified in the copyright notice associated with the article.
