Considerations for when reviewing human postmortem brain studies
The field of mental health research has largely shifted towards using in vitro and in vivo models of psychiatric disease, or clinically-based methods using peripheral tissues and neuroimaging approaches. Some researchers study molecules directly in the human brain, using tissues donated by people who lived with a psychiatric disorder. Over the last few years of studying the postmortem human brain, I’m finding it increasingly apparent that postmortem studies are in a class of their own – not that they are superior – but in that they need to be treated differently.
Postmortem brain studies are rare, precious, and valuable. For example, even when the results are negative, they are still informative. This is unlike other types of research such as animal or cellular models, where publishing negative results is difficult because the lack of significance is seen as a limitation of the model.* However, despite the value of postmortem human brain studies, they can sometimes be difficult to navigate through peer review.
A proportion of rejected postmortem manuscripts are likely a result of reviewers not being familiar with the strengths and limitations of postmortem brain analyses, whereby they place the same constraints on these works as they would with other types of research papers (animals/cell models/peripheral tissues/imaging etc.). This is just a short summary about why postmortem studies should be treated a little differently.
No scientific research is ever without limitations. Postmortem brain studies arguably have a few additional limitations. One of the reasons that postmortem studies are criticized is because there are multiple confounds which need to be considered. These include factors from the person’s life and clinical factors that can influence tissue integrity, including:
- How long the individual had the psychiatric disorder before they died (i.e. the years between the disease onset to their death)
- The individual’s agonal state (i.e. how, and how long, they suffered before they died)
- The postmortem interval (i.e. the time between when the person died and the brain was collected, autopsied and frozen for storage)
- How long the tissue has been stored in the freezer (even at minus 80 degrees Celsius, the tissue is still susceptible to degradation, but its minimal)
- The individual’s brain pH (this can often depend on the way the person died – for example, the suicide rate amongst individuals with schizophrenia is ~20%, and many of these suicides are by asphyxiation – dying this way can increase the brain’s acidity and influence molecular measurements).
- Their medication histories (since antipsychotics and antidepressants change brain chemistry), and the accuracy of this history due to compliance issues
- Whether the person smoked/did drugs/drank a lot (because these substances changes brain chemistry, too) – and the reliability of this information (i.e. was the patient really honest about reporting their substance abuse problems to their doctor?)
When brain banks acquire postmortem human brain tissues, they also check for tissue integrity to make sure it is suitable for conducting research. They do this by measuring the quality of the RNA in the tissue samples. RNAs are molecules that are particularly susceptible to degradation caused by factors such as those listed above, so they are a good indicator of whether the tissue is good enough to use for research. If the tissue quality is poor, the brains are excluded from the cohort so that only tissue that will produce reliable results is being examined.
Once we know tissue quality is good, we then want to match psychiatric cases as closely as possible with the brains of healthy people (the experimental controls). There are thus additional factors to consider: it is important that the cases and controls are matched as closely as possible for age, sex, and whether the samples were taken from the left or the right side of the brain – these are factors that influence our brain chemistry and molecular makeup, and can therefore impact on our results. Also because brain tissue is scarce, it can take quite a few years to put these cohorts together, so can be variability in the freezer storage time that can negatively influence the tissue integrity and impact on the results of a study.
Common practice is that we acquire good, standardised measures of these confounds across our cohorts – the pathologists/brain bank staff consolidate this information from the individual’s medical history – and we then factor these numbers into our statistical analyses as covariates. The outcome is that human brain tissues become very usable and valuable, despite the limitations. Further, new efforts to better address these confounds are constantly being made.
Source: Figure 2: Anatomical and histological views of the lesion and residual hippocampus, from ‘Postmortem examination of patient H.M.’s brain based on histological sectioning and digital 3D reconstruction.’ Nature Communications 2013.
* this is not to negate the usefulness of negative results: I am a big fan of publishing negative results in all fields, as I’ve discussed here.