There are anecdotal comments that every person with autism is autistic in their own way. Although autism is undoubtedly heterogeneous, a striking finding in brain reading studies of neurotypical people is the high degree of LY294002 commonality of neural representations of concepts across individuals. A classifier trained to identify the thoughts associated with physical objects like a banana from the neural activation patterns of a group of participants can then identify, with reasonable accuracy, the thoughts of a new participant whose data were not included in the training. This activation commonality probably arises because of the commonalities in the structure, function, and experience of human brains as they process information related to physical objects. But how would a psychiatric or neurological disorder affect the commonality among the members of the affected population, particularly in a domain of thought that is altered in the disorder? Given the apparent heterogeneity of autism, should there thus be less commonality among people with autism than among people without autism when they are thinking about social concepts? That is, if autism entails altered conceptions of social interactions, are the alterations heterogeneous across people with autism or is there a commonality? New machine learning methods allow a comparison of the commonality within the autism and the control groups. The central issue remains whether it is possible to identify a participant as autistic, not just on the basis of a fortuitous statistical relation, but on the basis of some fundamental alteration of the brain activity that underpins particular types of thought that are among the defining characteristic of the disorder. Below we first apply factor analysis to reduce the dimensionality of the brain activation evoked by the various social interactions. Then we perform classification of the multivoxel patterns that correspond to particular social interactions in order to identify the interaction and to distinguish the neural patterns of the two groups. The advantages of the approach are that it 1. focuses on the representations of social interactions, which are likely to be altered in autism and which like other concepts, are neurally represented by multiple voxels in multiple regions, and 2. is capable of detecting group differences in the activation patterns of multiple voxels in multiple regions. They are abundant and diverse from the poles to the equator, and from the deep sea to the intertidal. There is a rich fossil record dating back about 450 million years to the late Ordovician, and sea urchins have many unique body plan features and adaptations, including pentameral symmetry, a water vascular system, mutable collagen and larval budding and regeneration. Finally, there is a global fishery worth almost half a billion US dollars which motivates a deeper understanding of sea urchin biology. Sea urchins have long been a preferred laboratory organism due to the ease of obtaining large quantities of gametes, which can be fertilized externally and used in studies of fertilization, embryogenesis and larval development. It was primarily for these latter reasons.