Implicit Learning of Musical, but not Linguistically-Presented, Steady-State Grammars in ASD

Background: The structure of music is learned implicitly in infancy in much the same way as linguistic structure (grammar) is learned. Artificial implicit learning paradigms have demonstrated that both novel musical and novel linguistic structures can be learned quickly and implicitly by typically developing participants.  There is no manifest advantage in TD participants’ ability to learn the structures in either modality.
In autism, there is a marked preference and skill, explicit from infancy, for musical stimuli over linguistic stimuli. However, learning strategies for each modality have yet to be investigated in individuals with ASD.
Objectives: This experiment examines to what extent children with autism are able to show implicit learning for a grammar that is presented either musically or linguistically.
Methods: Using Reber’s Steady-State Grammar to construct a list of pseudo-word strings.  All grammatically correct strings were used in the training trial.  A five minute interval was allowed between training and testing to allow for consolidation of implicit knowledge. All novel stimuli were used in the testing trial, but half of the pseudo-word strings were grammatical and the other half were agrammatical.  Young people with autism were asked, in the test trial, to identify the familiar strings.  The experiment was then repeated with musical notes in the place of pseudo-words.
Results: Young people with autism were unable to identify the familiar pseudo-word strings with better-than-chance accuracy, but did achieve better-than-chance results when the stimuli were musical in nature. 
Conclusions: Young people with autism demonstrate implicit learning for grammar with musical, but not linguistic stimuli.