This section contains broad observations about the problems identified in my research, and ideas about how bio-informatics might assist. The goal of this section is to identify the top problems and guide my research based on these.
The biggest problems relate to chiari being a poorly understood disease, meaning that the cause(s) and “subtypes” have not been uncovered. Most of the literature is clinical in nature, either discussing the outcomes for a group of patients with some procedure, or attempting to review a particular type of chiari. Since chiari was thought to be a rare disease in the past 20 years, most of the research groups are first focusing on awareness and education before research.
While 4 “subtypes” have been identified (I,II,III,IV) with broad features, we still cannot say concretely that symptoms X –> Type Y. It is also apparent that the relationship between particular physiology and symptoms is not consistent. Some individuals might present with symptoms in absence of the salient physiological features of the disease (tonsils > 5mm below the foramen magnum, hydrocephalus / large ventricles, sometimes syrinx), while others might have the physiological features without any symptoms. Both physical features and symptoms vary between people, and it is not clear why the disease is comorbid with so many other disorders (developmental, connective tissue, hypermobility). It is also clear that some physiology that is considered normal (eg, hypermobility is common), alone is sufficient to contribute to the development of chiari. We also do not understand the differences between congenital and chiari acquired later in life as the result of some trauma. The primary two forms of treatment are the placement of a shunt (sometimes resolves symptoms and/or physical markers), or doing a decompression surgery to restore correct flow of CSF.
Diagnosis and treatment are problematic because physicians don’t have a clue why one treatment works for one person but not another. In many cases with complex symptoms it is not clear which symptoms are attributed to the chiari vs. something else, and in many of these cases patients do not get better and are forced to live with sensory & cognitive deficits and pain. I, however, do not agree with the literature and public awareness efforts that say that “we still don’t have a cure.” We do have procedures that are successful for some individuals. Individuals who claim the above are expressing the idea that “we don’t understand the circumstances of (some) particular chiari, and efforts to alleviate symptoms have not worked, and it has been expensive and terrible for the individual and the family.” Largely, correctly identified chiari will not kill an individual (as a rare cancer might), but it is usually the case that a diagnosis leads to many years of surgeries and distressing symptoms, and in the case that the chiari is not resolved, a lifetime of these symptoms. The reason that we cannot effectively treat it is because we do not understand it, and this is the rationale for research.
I originally was thinking of chiari as an independent disease that came in four types, and it might be an interesting problem to characterize each subtype based on physiological features. It is now my opinion that chiari is not an independent disease. Similar to autism, it is a label that is used to describe a heterogeneous group of physiology leading to somewhat similar symptoms. The problem can be reduced to an unbalanced environment housing the brain leading to sensory and motor deficit. Simply put, aberrant relationships between brain size, skull size and shape, and csf pressure lead to problems. The reason that the symptoms and physiology vary between individuals is because we likely are not dealing with the same underlying features that caused the disorder to begin with. The Monroe-Kellie hypothesis states that the skull cannot be compressed, and so the total volume of “stuff” inside (blood, csf, brain, and bone) is fixed.
I would hypothesize that the following types of features, either present at birth or elicited later in life as the result of trauma, can lead to some form of chiari:
The reasons for why the skull or brain are strangely shaped or sized would likely be related to some disorder or attributable to an individual genotype or phenotype (as the result of some gene, or group of physical features). For example:
It has been observed that chiari is comorbid with autism:
Autism (larger brain as result of aberrant pruning, BDNF gene?) –> less room in skull for csf and blood –> development of syrinx and herniation of tonsils
It has been observed that chiari is comorbid with structural abnormalities:
Smaller foramen magnum or fused bones –> disrupted CSF flow –> development of syrinx / herniation
It has been observed that chiari is comorbid with connectivity disorder/hypermobility:
Inflammation of ligament in front of cord –> less space –> further herniation/syrinx clival-axial angle less than 150 degrees –> same
If we largely have different types of problems with skull size/shape, brain size, range of movement, could it be that long term varying distribution of CSF pressure as a result of these physiological features leads to different symptoms or even cognitive or motor deficits associated with a psychological disorder? Could it be the case that if we measured these physical features across a broad range of disorders, we would uncover new patterns of physiological features associated with the disorder?
I am having a hard time focusing this idea concretely around chiari. Can this problem broadly be thought of as an investigation into how differences in brain size and skull morphometry can lead to symtomatology and functional/behavioral deficit? What is more interesting is the effort to use publicly available imaging data to build a model of the house that the brain lives in, extracting features from this space, and then looking for features associated with disorder. Most neuroscientists focus on structure or function of the brain removed from csf and skull, and I think that more work should be done to study the big picture. Further, this research is done by shoving individual brains into a standard space when it is really the case that everyone has different morphometry. This variation should be explored, because if a particular area of the brain has aberrant function or structure in some clinical group, could it be because higher csf pressure / bony abnormality has slowly caused apoptosis of neurons?
In a sense, chiari can be thought of as an extreme case of some of these features, a case for which we know that aberrant structure –> problem. This could be useful for an early model or hypothesis that would aim to identify clusters of physical features associated with particular symptoms. For example, I would want to compare structural markers, and a csf gradient map between normal and chiari patients, and identify differences between healthy control and chiari. I could also look at gender differences, and look for associations with other meta data (behavioral, clinical, fMRI, DTI, etc). I would then want to explore how these same features are expressed across different disorders, and look for correlations with this same meta data. For example, let’s say that I find that features X, Y, and Z are significantly different in chiari patients as compared to healthy control, and manifested with some particular symptom(s). I might first use these finding to better identify subtypes of chiari not as I,II,III, or IV, but rather as groups of well defined features, and then move to other disorders (below).
When there is a link between features –> symptoms, I then might ask if we can find similar relationships in other disorders. I would calculate the distribution of these features in an ADHD vs HC population, and look for correlation between structural and functional deficit identified in the literature with these distributions. It would be desirable to build a complete model of structural patterns that are linked to symptoms, and then attempt to predict symptoms based on extracted features and other associated meta data from a novel case. Across different psychopathology, it would be useful to be able to say “these cognitive deficits that were associated with your label of “ADHD” are likely caused by a small foramen magnum and fused c2.” or “The autism is being exacerbated by features X,Y,Z leading to a higher intracranial pressure, meaning there is less blood flow to these regions, and so we can help your child by fixing the structural deficit.” The discovery might be that much of the aberrant symptoms that are typical of other disorders actually have very simple explanations. The large effort in neuroscience that only looks at brain structure and function is akin to studying a person without looking at his or her environment. Most researchers throw out the skull and csf data that is valuable for understanding the whole picture.
It is not clear how I might clearly validate my results, but I might look at genes involved with bone / skull formation… need to think more about this!