Faculty Bibliography

BACKGROUND:Abstraction of critical data from unstructured radiologic reports using natural language processing (NLP) is a powerful tool to automate the detection of important clinical features and enhance research efforts. We present a set of NLP approaches to identify critical findings in patients with acute ischemic stroke from radiology reports of computed tomography (CT) and magnetic resonance imaging (MRI). METHODS:We trained machine learning classifiers to identify categorical outcomes of edema, midline shift (MLS), hemorrhagic transformation, and parenchymal hematoma, as well as rule-based systems (RBS) to identify intraventricular hemorrhage (IVH) and continuous MLS measurements within CT/MRI reports. Using a derivation cohort of 2289 reports from 550 individuals with acute middle cerebral artery territory ischemic strokes, we externally validated our models on reports from a separate institution as well as from patients with ischemic strokes in any vascular territory. RESULTS:In all data sets, a deep neural network with pretrained biomedical word embeddings (BioClinicalBERT) achieved the highest discrimination performance for binary prediction of edema (area under precision recall curve [AUPRC] > 0.94), MLS (AUPRC > 0.98), hemorrhagic conversion (AUPRC > 0.89), and parenchymal hematoma (AUPRC > 0.76). BioClinicalBERT outperformed lasso regression (p < 0.001) for all outcomes except parenchymal hematoma (p = 0.755). Tailored RBS for IVH and continuous MLS outperformed BioClinicalBERT (p < 0.001) and linear regression, respectively (p < 0.001). CONCLUSIONS:Our study demonstrates robust performance and external validity of a core NLP tool kit for identifying both categorical and continuous outcomes of ischemic stroke from unstructured radiographic text data. Medically tailored NLP methods have multiple important big data applications, including scalable electronic phenotyping, augmentation of clinical risk prediction models, and facilitation of automatic alert systems in the hospital setting.

Attracted by the appealing advantages of optogenetics, many nonhuman primate labs are attempting to incorporate this technique in their experiments. Despite some reported successes by a few groups, many still find it difficult to develop a reliable way to transduce cells in the monkey brain and subsequently monitor light-induced neuronal activity. Here, we describe a methodology that we have developed and successfully deployed on a regular basis with multiple monkeys. All devices and accessories are easy to obtain and results using these have been proven to be highly replicable. We developed the "in-chair" viral injection system and used tapered and thinner fibers for optical stimulation, which significantly improved the efficacy and reduced tissue damage. With these methods, we have successfully transduced cells in multiple monkeys in both deep and shallow cortical areas. We could reliably obtain neural modulation for months after injection, and no light-induced artifacts were observed during recordings. Further experiments using these methods have shown that optogenetic stimulation can be used to bias spatial attention in a visual choice discrimination task in a way comparable to electrical microstimulation, which demonstrates the potential use of our methods in both fundamental research and clinical applications.

Neuroprosthesis research aims to enable communication between the brain and external assistive devices while restoring lost functionality such as occurs from stroke, spinal cord injury or neurodegenerative diseases. In future closed-loop sensorimotor prostheses, one approach is to use neuromodulation as direct stimulus to the brain to compensate for a lost sensory function and help the brain to integrate relevant information for commanding external devices via, e.g. movement intention. Current neuromodulation techniques rely mainly of electrical stimulation. Here we focus specifically on the question of eliciting a biomimetically relevant sense of touch by direct stimulus of the somatosensory cortex by introducing optogenetic techniques as an alternative to electrical stimulation. We demonstrate that light activated opsins can be introduced to target neurons in the somatosensory cortex of non-human primates and be optically activated to create a reliably detected sensation which the animal learns to interpret as a tactile sensation localized within the hand. The accomplishment highlighted here shows how optical stimulation of a relatively small group of mostly excitatory somatosensory neurons in the nonhuman primate brain is sufficient for eliciting a useful sensation from data acquired by simultaneous electrophysiology and from behavioral metrics. In this first report to date on optically neuromodulated behavior in the somatosensory cortex of nonhuman primates we do not yet dissect the details of the sensation the animals exerience or contrast it to those evoked by electrical stimulation, issues of considerable future interest.

Methods on rendering neurons in the central nervous system to be light responsive has led to a boom in using optical neuromodulation as a new approach for controlling brain states and understanding neural circuits. In addition to the developing versatility to "optogenetically" labeling of neural cells and their subtypes by microbiological methods, parallel efforts are under way to design and implement optoelectronic devices to achieve simultaneous optical neuromodulation and electrophysiological recording with high spatial and temporal resolution. Such new device-based technologies need to be developed for full exploitation of the promise of optogenetics. In this paper we present single- and multi-element optoelectronic devices developed in our laboratories. The single-unit element, namely the coaxial optrode, was utilized to characterize the neural responses in optogenetically modified rodent and primate models. Furthermore, the multi-element device, integrating the optrode with a 6×6 microelectrode array, was used to characterize the spatiotemporal spread of neural activity in response to single-site optical stimulation in freely moving rats. We suggest that the particular approaches we employed can lead to the emergence of methods where spatio-temporal optical modulation is integrated with real-time read out from neural populations.