I am curious why DSM-V says autism spectrum disorder and schizophrenia spectrum disorder are compatible. I am geologist, and all that topic about autism smells to me like a biologic cold adaptation with a sligthy different origin than schizophrenia.

Just I am asking, as I should be missing something, but:

  • In this paper it is said autists have low CB1 receptors density.
  • In this other paper it is said schizophrenics have high CB1 receptors density.

Why the compatibility then? CB1 receptors are not quantum particles.


1 Answer 1


We do not understand what causes either autism (ASD) or schizophrenia in the brain. There are some theories and ideas, but nothing conclusive and certainly nothing as simple as a single receptor. There may be gene variants that are associated with ASD or schizophrenia, or findings in histology or neuroimaging that correlate with ASD or schizophrenia.

In the first paper that you link, about CB1 receptor density in autism, one of the citations, reference 21, is this one:

Purcell, A. E., Jeon, O. H., Zimmerman, A. W., Blue, M. E., & Pevsner, J. (2001). Postmortem brain abnormalities of the glutamate neurotransmitter system in autism. Neurology, 57(9), 1618-1628.

This paper includes two different genome-wide studies, each with very few individuals. For the CB1 receptor specifically, one of the studies showed more CB1 and the other showed less. It isn't clear to me why the authors have taken this to mean "CB1 is low in ASD"; the evidence they provide for this is extremely weak and perhaps non-existent, at least based on the given reference. You absolutely should not take this to mean "low CB1 causes autism" or that everyone with ASD has less CB1 than everyone without ASD.

In the second paper, about schizophrenia, there are several statements that suggest higher CB1 receptor density in various brain regions, but also examples of failures to find any differences. Again, it seems there is some evidence for this but it is weak and not necessarily causal. They also mention caveats in interpreting these data:

Assuredly, there are many confounding factors influencing the results of post-mortem studies, such as antipsychotic treatment, cannabis smoking and the heterogeneity of biochemical techniques.

Let's look into one of these papers results, Wong, D. F., Kuwabara, H., Horti, A. G., Raymont, V., Brasic, J., Guevara, M., ... & Cascella, N. (2010). Quantification of cerebral cannabinoid receptors subtype 1 (CB1) in healthy subjects and schizophrenia by the novel PET radioligand [11C] OMAR. Neuroimage, 52(4), 1505-1513.. Here is a figure from that paper (figure 5):

Wong et al Figure 5: PET imaging by age and schizophrenia

In this figure, there are 4 panels each corresponding to a different brain region. The y-axis is a measure of CB1 density (after some processing); the x-axis is age. The black symbols, solid lines, and dashed lines represent non-schizophrenic participants, the average linear relationship with age, and confidence intervals around that relationship, respectively. You could roughly consider any points between the dashed lines as "similar to normal for a given age". Notice where the red points are, indicating schizophrenia: for most brain regions and most participants, there is no real way to distinguish them from "normal" using this measure. There might be an overall group difference between schizophrenia and controls, but this is mostly due to a couple of outliers. Many of the schizophrenia participants have lower CB1 than the average non-schizophrenia patient: all the red points below the black lines. In this particular paper, they compared 19 brain regions and only found a significant difference (p<0.05) in one of them, the Pons, which did not appear to be a place they expected to find a difference ahead of time. It does not appear the authors corrected for multiple comparisons. You would expect to find a significant result p<0.05 purely by chance in one out of 20 comparisons - that's what the 0.05 threshold means, that 5% of the time an effect that large occurs if you assume there is no actual difference; it's unlikely this result would survive correction so it should not be considered significant. They need a much larger sample to assess if there is any real difference.

That doesn't mean that there isn't some real group difference here when considering the other studies as well, but a difference between groups is not the same as saying that every individual in each group follows the same pattern, and like for ASD the evidence for a link here seems very very weak.

In summary, the relationship between CB1 density and ASD or schizophrenia is quite weak and does not support a causal link between that receptor either condition. From the available data, it seems that most people with either ASD or schizophrenia do not have CB1 receptor densities that distinguish from the overall population distribution.

  • $\begingroup$ Hello Bryan. At the hospitallization I earned rep with the psychiatrists and some of them are going to read this and my theory of autism, my phd subjet in a couple of years, tomorrow. As I told you in the chatroom, not convinced. Some of the schizophrenics with not high CB1 could be delusional patients and the autist of that ancient post mortem study schizophrenics. Rats with induced autism with valproate show low CB1 in recent papers. Recent papers sugest also delusional disorder has mild negative symptomatology $\endgroup$
    – user33884
    Feb 6 at 18:47
  • $\begingroup$ @Universal_learner I'm happy to review other specific papers about this; the evidence presented in the ones you mentioned in the original post is all I've evaluated. That evidence is very weak. It may still be of interest to study, but I wouldn't base conclusions on such weak evidence. $\endgroup$
    – Bryan Krause
    Feb 6 at 18:57

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service and acknowledge you have read our privacy policy.