3 ms·
Two samples each with n = 13. ASD is widely acknowledged to be a highly heterogeneous condition which is likely to have multiple independent causes, and which
by skissane 5y ago
Two samples each with n = 13.
ASD is widely acknowledged to be a highly heterogeneous condition which is likely to have multiple independent causes, and which has unclear boundaries with other diagnoses (especially ADHD)
So a study which effectively treats ASD as a single undifferentiated condition, and has rather small sample sizes too, I think should be treated with a great deal of scepticism.
- 2T1Qka0rEiPr 5y agoYes, I read down to this point and stopped. What significance can you really draw from a sample size of 26? (Serious question to academics out there)
- renewiltord 5y agoTo answer your serious question, n=26 is more than sufficient depending on effect size and sampling method. Think about it, if you have 26 people randomly sampled and you give half a pill and the other half a placebo and the first half dies moments after ingestion, how confident are you that the pill is poisonous? If the effect size is small, you may fail to detect it with a small population. If the sampling is biased, then you’re going to have a problem even if you have a massive population studied.
- tpoacher 5y agoThe "small sample size" has been drilled to most people as a knee jerk red flag indicating a statistically unreliable conclusion, but this is not always the case. Depending on the desired outcome, a small sample size may actually increase the reliability of the result. This seems like one of those cases. Think of it like this. The aim is to show a statistically significant difference. With a large enough sample, you will eventually show significance (hence Fisher's famous anecdote of "get more data"), but the effect size may be trivial. Whereas, with a small sample, only non-trivial effect sizes will achieve significance, and therefore achieving significance with such a small sample tells you something about the nontrivial extent of the effect size. So you could argue that the extent of methylation change was so large, that it achieved significance in a sample even as small as n=13!
- nicoburns 5y agoI agree with heterogeneous. But why do you think that auggests multiple causes. Commonalities across the neurodivergent spectrum intuitively lead me to suspect the opposite, that there is some common mechanism underlying all of them.
- skissane 5y agoBecause the same symptoms can be produced by very different brain structures. Have you read this paper: https://www.nature.com/articles/s41398-019-0631-2 https://www.nature.com/articles/s41398-019-0631-2 They find that the four-way distinction between ASD, ADHD, OCD and "typically developing" has poor correspondence to the observed interaction between cortical morphology and behavioural symptoms. If you take their clusters – https://www.nature.com/articles/s41398-019-0631-2/figures/3 https://www.nature.com/articles/s41398-019-0631-2/figures/3 – as "what is really going on in the brain", then the same "what is really going on in the brain" can produce all four of ASD, ADHD, OCD and typical development, at different proportions (clusters 1-5). Whereas, conversely, there is a subgroup of ASD (cluster 10) which doesn't overlap with ADHD or OCD or typicality. I'd also recommend reading https://link.springer.com/article/10.1007/s40489-016-0085-x https://link.springer.com/article/10.1007/s40489-016-0085-x
- cjohansson 5y agoI would turn the argument around and say that what we usually diagnose as ADHD or ASD are infact many different brain variations. Having for example ADHD doesn't mean your brain has any similiarity to another ADHD brain. The diagnosis does not involve physical brain measurements in any way
- skissane 5y agoIf the boundaries between neurodevelopmental disorders lack validity – certainly they lack biological validity, and it has been argued [0] that they lack construct validity as well – maybe we should respond by merging them? Merge ADHD, ASD, etc, into one single overarching disorder? That is one way of interpreting Christopher Gillberg's ESSENCE proposal. It is also, at least somewhat, the recommendation of Lynn Waterhouse's book Rethinking Autism: Variation and Complexity, which ends with the proposal to replace ASD with the much broader concept of "phenotypes of neurodevelopmental social impairment" [0] See https://doi.org/10.1002/aur.1832 https://doi.org/10.1002/aur.1832 in particular the authors' claim that "our review of research has led us to conclude that using the ASD diagnosis in basic and translational research cannot provide valid conclusions because the diagnosis lacks both biological and construct validity" and also "The ASD diagnosis lacks boundary construct validity..."