Browsing by Subject "Causality"

Causal inference methods are commonly used across domains to aid high-stakes decision-making. The validity of causal studies often relies on strong assumptions that might not be realistic in high-stakes scenarios. Inferences based on incorrect assumptions frequently result in sub-optimal decisions with high penalties and long-term consequences. Unlike prediction or machine learning methods, it is particularly challenging to evaluate the performance of causal methods using just the observed data because the ground truth causal effects are missing for all units. My research presents frameworks to enable validation of causal inference methods in one of the following three ways: (i) auditing the estimation procedure by a domain expert, (ii) studying the performance using synthetic data, and (iii) using placebo tests to identify biases. This work enables decision-makers to reason about the validity of the estimation procedure by thinking carefully about the underlying assumptions. Our Learning-to-Match framework is an auditable-and-accurate approach that learns an optimal distance metric for estimating heterogeneous treatment effects. We augment Learning-to-Match framework with pharmacological mechanistic knowledge to study the long-term effects of untreated seizure-like brain activities in critically ill patients. Here, the auditability of the estimator allowed neurologists to qualitatively validate the analysis via a chart-review. We also propose Credence, a synthetic data based framework to validate causal inference methods. Credence simulates data that is stochastically indistinguishable from the observed data while allowing for user-designed treatment effects and selection biases. We demonstrate Credence's ability to accurately assess the relative performance of causal estimation techniques in an extensive simulation study and two real-world data applications. We also discuss an approach to combines experimental and observational studies. Our approach provides a principled approach to test for the violations of no-unobserved confounder assumption and estimate treatment effects under this violation.

Background

Methods

Results

Conclusion

Substantial correlative evidence links the synchronized, oscillatory neural firing patterns that emerge in Parkinson’s disease (PD) in the frequency range of 13-30Hz (termed “beta band”) with the development of bradykinesia and akinesia. Yet, a causal link between these beta frequency oscillations and symptoms of bradykinesia has not been demonstrated. I tested the hypothesis that the synchronized beta oscillations that emerge in PD are causal of symptoms of bradykinesia/akinesia through studies in intact and parkinsonian animals as well as PD patients.

Regarding the rat studies, I designed novel stimulation patterns to mimic the temporal characteristics of the beta oscillatory bursting pattern seen in single units in PD rats and patients. I applied these beta frequency patterned stimulus trains along with continuous frequency controls over a range of amplitudes via stimulating electrodes implanted unilaterally into the subthalamic nucleus (STN) of healthy and PD rats and assessed the effects on unit activity in the substantia nigra reticulata (SNr) and performance in motor tasks designed to assess forelimb bradykinesia and gross locomotor activity. I quantified the degree of unit entrainment in the SNr as a function of pattern and amplitude by calculation of the excitatory effective pulse fraction (eEPF) [1]. I further quantified the increase in SNr unit spectral beta frequency power due to the applied stimulation paradigms. I found that the beta-patterned paradigms were superior to low frequency controls at entrainment and induction of beta power in downstream substantia nigra reticulata (SNr) neurons. However, I found no deleterious effects on motor performance across a wide battery of validated behavioral tasks.

In PD patients, my objective was to determine how beta frequency oscillatory activity varies with disease progression and severity in human PD patients using cortical electroencephalogram (EEG). I recorded EEG in twenty-five PD patients of varying disease severity after overnight abstinence from PD medication. I recorded EEG at rest (eyes open and closed) and while patients performed various hand motor tasks. These tasks included one-handed isometric grip and rapid open/close movements. For each EEG channel data-stream for each patient, I calculated the total percent of spectral power in the beta band (PSP-β) with respect to movement state. I used stepwise regression to predict UPDRSIII scores from the normalized PSP-β values calculated for each channel during the ‘Eyes Open Rest’ state and found a significant, predictive regression equation. I assessed the relationship between UPDRSIII score and cortical coherence using linear regression and found significant, positive correlations between UPDRSIII score and coherence at beta band frequencies between pre-motor-motor and motor-somatosensory cortical areas. I observed phase amplitude coupling (PAC) between beta and gamma (30-200Hz) frequencies at rest and found it to be significantly altered by task, but found no effect of motor symptom progression on mean PAC.

My data suggest that certain metrics of beta band activity in pre-motor, motor, and somatosensory brain regions at rest may serve as a marker for degree of motor impairment, but that beta frequency oscillations may be an epiphenomenon and not necessary or sufficient for the generation of bradykinesia/akinesia in PD.