Friday, June 1, 2018

Structural Development of the Human Brain - The Clinical View

Last year I published an article in the excellent journal Human Brain Mapping that analyzes a huge population of patients across a wide range of developmental ages who received clinical imaging at Boston Children's Hospital. You can check out the article here.


This research project represents the first phase of a rather different approach to neuroimaging research than is most commonly practiced. A typical magnetic resonance imaging (MRI) study of the brain involves what's called a prospective study design wherein researchers recruit patients to be included in their study by having physicians ask their patients to participate, by posting advertisements in clinical settings and by offering participants incentives (usually a small compensatory payment in exchange for participation), etc. My study is quite different in that we reviewed an enormous number of patient medical records in order to identify those clinical patients who are most likely to have typically developing brains (a hospital radiologist indicated that their MRI was normal, their medical records didn't indicate neurological disorders, etc.). By structuring our study this way, we were able to collect a very large pool of patients to include in our analysis (993 examinations). Large sample sizes help to overcome some of the limitations of existing studies and large quantities of clinical data can be used to validate findings from traditional prospective studies. 

This work is foundational towards future research assessing a variety of medical conditions for which accumulating many samples is extremely challenging. For instance, in autism (and other potentially serious developmental conditions), a prospective study design is very time consuming and extremely expensive to implement. While retrospective clinical review of data (as per my study) necessarily samples patients with additional medical conditions (the patient needed a reason to get the MRI), prospective study designs are not likely to contain a representative sample of the most severe forms of the medical condition being studied (caregivers are less likely to participate in a study when it involves extra unnecessary hospital visits, subjecting the patient to noisy MRI exams, etc.). This clinical dataset in my accompanying article represents our best attempt at discerning who among our hospital's clinical patients are truly typically developing (i.e. healthy). This dataset can be used to study healthy brain development, but also to act as controls for comparative analyses with patients with autism, attention deficit hyperactivity disorder, etc. Work on these conditions is under development or under review for publication.

This study also included measurements of hemispheric asymmetry (left-right imbalances in the brain) as well as demonstrating that these methods can be extended to the regional assessment of brain maturation as is illustrated in the figure below which can be found in the original article.