Analysis: Several categories of measures that we collect have the potential to differentiate particular aspects of the centrally mediated impact that an individual’s neurological condition has on their information processing capacity.
For example, examination of the adaptation metrics, in case studies (figure above) in which metrics were obtained before and after treatment and/or recovery demonstrates a potential utility for these types of measure. Two subjects were given a pharmacological treatment (vulvodynia and autism patients were administered a GABA agonist and citalopram, respectively). The patient with alcoholism was tested at initial sobriety and after twelve weeks of sobriety (during which time he was also administered baclofen, a GABAb agonist). The concussed subject data is a comparison of 3 day and 7 day post--concussion (no drugs administered). Thus, observations such as this one – even though it is non-specific (appears to be sensitive to disruptions caused by a wide spectrum of disorders), have the potential for aiding clinicians with diagnosis and assessing treatment efficacy of multiple neurological disorders.
Multiparametric analysis: Combining all metrics obtained from each subject creates a sensory profile that can be assessed with mathematical tools such as PCA (principal component analysis) and SVM (support vector machine; a machine learning algorithm). SVM will be trained via clinical diagnostics (obtained from the program project) to identify the markers that are most sensitive to the different pain categories as well as the sub-categories within each pain group. Sample PCA plots are shown at the right to demonstrate potential utility in segregating patient groups. A multiparametric test (typical observations result in 9-12 parameters that yield the same number of principal components) will often result in a PCA plot showing some overlap in 2D (first two PCs), but segregation is more easily observed in 3D. SVM utilizes all parameters in addition to clinical diagnostics to learn where segregation of data and/or categorization of groups occurs. Data mining of the generated database, which will include clinical assessment of patients, medication history and treatment outcomes will be utilized to optimize parameters for different patient categories and sub-categories.