OK I'm a little confused right now.
Not so long ago I talked about the paper from Horvath and colleagues  (see here) concluding that "the limited data currently available suggest that ω-3 FA [omega-3 fatty acid] supplementation does not enhance the performance of children with ASD [autism spectrum disorder]." Such a conclusion was based on the application of a systematic review and meta-analysis of the available peer-reviewed literature up to August 2016.
Now however, another systematic review and meta-analysis on the topic has emerged from Mazahery and colleagues  and reported something a little bit different: "Populations with ASD have lower n-3 LCPUFA status and n-3 LCPUFA supplementation can potentially improve some ASD symptoms." Don't you just love science!
OK, so what could be the reason(s) for the differing conclusions reached by the reviews on this topic? Well, Mazahery and colleagues (the most recent review) actually conducted two meta-analyses: "meta-analysis 1 compared blood levels of LCPUFA and their ratios arachidonic acid (ARA) to docosahexaenoic acid (DHA), ARA to eicosapentaenoic acid (EPA), or total n-6 to total n-3 LCPUFA in ASD to those of typically developing individuals (with no neurodevelopmental disorders), and meta-analysis 2 compared the effects of n-3 LCPUFA supplementation to placebo on symptoms of ASD." Horvath et al only conducted one meta-analysis in their study roughly equivalent to meta-analysis 2 presented by Mazahery looking at the effects of fatty acid supplementation on the presentation of autism. Mazahery and colleagues also surveyed the literature up to May 2016 and found four randomised-controlled trials (RCTs) (N=107) whereas Horvath et al (who published earlier!) surveyed the literature up to August 2016 and found five RCTs (N=183). Indeed, it appears that based on that last 'difference' one might see how the grand 'top of the scientific hierarchy' meta-analysis is yet again, only as good as the data it contains. And a certain celebrity in science circles seems to agree...
Where next I hear you ask? Well, I'd be tempted to follow the recommendations of Mazahery and colleagues when they suggest that: "Further research with large sample size and adequate study duration is warranted to confirm the efficacy of n-3 LCPUFA." Indeed, there are already studies to watch in this area. That and recognising that within the vast plurality that is the autisms it is not totally outside the realms of possibility that specific parts of the autism spectrum might be more vulnerable to fatty acid issues than others. Oh, and don't forget that outside of impacting autistic 'performance' (or not), fatty acid supplementation does seem to have other health-related properties too...
 Horvath A. et al. ω-3 Fatty Acid Supplementation Does Not Affect Autism Spectrum Disorder in Children: A Systematic Review and Meta-Analysis. J Nutr. 2017 Jan 11. pii: jn242354.
 Mazahery H. et al. Relationship between Long Chain n-3 Polyunsaturated Fatty Acids and Autism Spectrum Disorder: Systematic Review and Meta-Analysis of Case-Control and Randomised Controlled Trials. Nutrients. 2017 Feb 19;9(2). pii: E155.
Mazahery H, Stonehouse W, Delshad M, Kruger MC, Conlon CA, Beck KL, & von Hurst PR (2017). Relationship between Long Chain n-3 Polyunsaturated Fatty Acids and Autism Spectrum Disorder: Systematic Review and Meta-Analysis of Case-Control and Randomised Controlled Trials. Nutrients, 9 (2) PMID: 28218722