Eight Types of Schizophrenia? Not so Fast…

Schizophrenia has been widely researched for many years now, yet still remains a big question mark in the realm of mental health. In 2014; however, Arnedo et al. published a study stating that schizophrenia is a heterogeneous group of disorders. 

The study received a lot of attention and one of the most pressed upon topics of the paper was that there were eight subtypes of schizophrenia under the umbrella term. However, there is simply not enough data within the study to back up their claims. 

Instead, the paper focuses on analyses that supposedly link clusters of data points to patterns of data. We believe that Arnedo et al. has found the most complex method of understanding these links rather than the simpler explanations. 

Let’s take a look at a few unanswered questions and concerns that we have with the original study. 

Eight types of schizophrenia Not so fast…

The X Chromosome

Arnedo et al. believed that the chrX was a focus on the subtypes of schizophrenia. However, there was nowhere in the paper do they explain how they have factored sex into the matrix factorization. 

As men will only have one chrX, the male genotypes can only be 0 or 1. However, females can have two copies of chrX. This means that the genotypes for females will be either 0, 1, or 2. 

So, instead of the chrX being a clinical feature of schizophrenia, it might be down to the sex of the participants instead.  


Repeating a study is a good sign of validity and therefore generating confidence in their findings. While Arnedo et al. did replicate their study and have two samples to view, it is unclear what they replicated and how they did it.

Due to this, we are skeptical about the replication process and the validity of the study as a whole. 


Arnedo et al. used three samples of participants, two of which were mainly European or African. The third sample was from southern Europe only. It is a well-known fact that ancestry plays a big part in genetic studies, so it has to be adjusted to allow for different ancestries. 

Table S3 showed significant differences between a European and African group. The mean for the difference in frequency was 0.27 for this table, while 0.19 was used for all other SNPs. This significant difference sparked out concerns about potential bias in population stratification. 


Arnedo et al. has made extraordinary claims with this study, leaving it open for many interpretations and concerns. There is simply not enough evidence, in our opinion, for them to be making such bold claims. 

They have picked and chosen what to showcase in their studies, such as the replication process and ancestry. We also have concerns about gender playing a role in the clustering analysis and being mistaken for symptoms of schizophrenia. 

Overall, we think that Arnedo et al. study is interesting and thought-provoking, but we would need significantly more research to prove that it was actually relevant to schizophrenia. Be careful before you interpret the results.