A Novel Way to Study Mental Illness: Dimensional Analysis
The research approach adopted by the JBRF consortium of investigators is called a dimensional analysis. This approach is in contrast to the categorical approach which has been in use since the 1980′s (click here to read about the categorical approach).
Dimensional analysis makes use of advances in computing which now allow massive amounts of variable data to be subjected to statistical analysis. The analysis is therefore able to incorporate many symptoms and/or behaviors and, importantly, to include degree-of-severity measures for those symptoms and behaviors. Previously they were only considered to be present or absent. A dimensional analysis can reveal;
- for the group of people gathered for the study, and
- the symptoms and behaviors selected for inclusion in the study,
- which symptoms and/or behaviors cluster together more often with each other than they do with others,
- for the people in that group.
The clusters that result from this process can be thought of as dimensions of behavior.
One assumption is that the symptoms and behaviors form clusters because, when the brain acts to produce one of them, the others are closely associated. They belong together as a unified expression and should be seen as a single entity rather than split into different disorders.
If the analysis net has been cast wide enough to allow all of the symptoms that actually associate with each other to do so, and if the analysis is repeated again and again, all the while refining the group of individuals whose symptom-profiles are included in the analysis, then eventually a meaningful profile of behavioral dimensions may emerge. This profile can lead investigators to a novel and coherent view of a syndrome that occurs in a large group of individuals. Further, this view may more directly reflect the source of an illness than the consensus-driven symptom lists that underlie today’s classifications.
The leap from dimensional profile to the source of an illness is possible because, like the advances in computing that make a dimensional analysis possible, our knowledge of brain areas and their functions has also advanced tremendously. There is now a much better, albeit incomplete, understanding of which parts of the brain are involved with which kinds of behavior and how they do it. This allows us to relate the dimensions of behavior to the those brain areas likely to be involved. If the analysis has pursued the right track, eventually an entire neuroanatomical hypothesis can emerge that more closely fits the natural presentation of the illness.
That has been the experience of the JBRF consortium investigators. A dimensional symptom profile emerged that led them to consider certain brain areas. Further research revealed the pathways that connect those areas. Soon after that, a complete and robust hypothesis evolved. The hypothesis can account for the full range of symptoms experienced by the children they have studied. The profile brings together into a single condition, symptoms that were never before directly associated with each other, and some that were never before associated with conditions defined as psychiatric illnesses.
Importantly, as in all of medicine, the identification of the pathophysiology that underlies an illness, dramatically improves the likelihood of finding effective treatments.