Justin Baker

The performance of machine learning algorithms used for neural decoding of dexterous tasks may be impeded due to problems arising when dealing with high-dimensional data. The objective of feature selection algorithms is to choose a near-optimal subset of features from the original feature space to improve the performance of the decoding algorithm. The aim of our study was to compare the effects of four feature selection techniques, Wilcoxon signed-rank test, Relative Importance, Principal… Show more

The performance of machine learning algorithms used for neural decoding of dexterous tasks may be impeded due to problems arising when dealing with high-dimensional data. The objective of feature selection algorithms is to choose a near-optimal subset of features from the original feature space to improve the performance of the decoding algorithm. The aim of our study was to compare the effects of four feature selection techniques, Wilcoxon signed-rank test, Relative Importance, Principal Component Analysis (PCA), and Mutual Information Maximization on SVM classification performance for a dexterous decoding task.

A nonhuman primate was trained to perform small coordinated movements—similar to typing. An array of microelectrodes was implanted in the hand area of the motor cortex of the NHP and used to record action potentials (AP) during finger movements. A Support Vector Machine was used to classify which finger movement the NHP was making based upon AP firing rates. We used the SVM classification to examine the functional parameters of robustness to simulated failure and longevity of classification. We also compared the effect of using isolated-neuron and multi-unit firing rates as the feature vector supplied to the SVM.

Main results: The average decoding accuracy for multi-unit features and single-unit features using Mutual Information Maximization across 47 sessions was 96.74 ±3.5% and 97.65 ±3.36% respectively. The reduction in decoding accuracy between using 100% of the features and 10% of features based on MIM was 45.56% and 4.75%for multi-unit and single-unit features respectively. MIM had best performance compared to other feature selection methods.

These results suggest improved decoding performance can be achieved by using optimally selected features. The results based on clinically relevant performance metrics also suggest that the decoding algorithm can be made robust by using optimal features and feature selection algorithms. Show less

We hypothesized that a novel surgical method, in which three-dimensional (3-D) preoperative planning software creates a patient-specific surgical model that is used with a reusable and adjustable tool, could substantially improve the positioning accuracy of the glenoid guide pin used in TSA. We tested this method using bone models from patients with shoulder pathology and compared the results with those achieved using standard of care surgical methods.

Three surgeons with varied surgical… Show more

We hypothesized that a novel surgical method, in which three-dimensional (3-D) preoperative planning software creates a patient-specific surgical model that is used with a reusable and adjustable tool, could substantially improve the positioning accuracy of the glenoid guide pin used in TSA. We tested this method using bone models from patients with shoulder pathology and compared the results with those achieved using standard of care surgical methods.

Three surgeons with varied surgical experience placed a guide pin in nine bone models from patients with a variety of glenohumeral arthritis severity using (1) standard instrumentation alone, (2) standard instrumentation and 3-D preoperative surgical planning, and (3) the reusable transfer device and 3-D preoperative surgical planning. A postoperative 3-D CT scan of the bone model was made and registered to the preoperative plan, and the differences between the actual and planned pin locations and trajectories were measured.

Use of the standard instrumentation combined with 3-D preop planning software improved guide pin positioning compared with standard instrumentation and preop planning using 2-D imaging. The accuracy of pin positioning increased by 4.5° ± 1.0° in version (p< 0.001), 3.3° ± 1.3° in inclination (p = 0.013), and 0.4 ± 0.2 mm in location (p = 0.042). Use of the adjustable, reusable device and the 3-D software improved pin positioning by a further 3.7° ± 0.9° in version, 8.1° ± 1.2° in inclination, and 1.2 ± 0.2 mm in location (p < 0.001 for all) compared with standard instrumentation and the 3-D software; the improvement compared with use of standard instrumentation with 2-D imaging was 8.2° ± 0.9° in version, 11.4° ± 1.2° in inclination, and 1.7 ± 0.2 mm in location (p < 0.001 for all).

Use of 3-D preoperative planning and use of the patient-specific bone model and transfer device both improved the positioning accuracy of the pin used to guide placement of the glenoid component in TSA. Show less

Dexterous finger movements can be decoded from neuronal action potentials acquired from a non-human primate using a chronically implanted Utah Electrode Array. We have developed an algorithm that can, after training, detect and classify individual and combined finger movements without any a priori knowledge of the data, task, or behavior. The algorithm is based on changes in the firing rates of individual neurons that are tuned for one or more finger movement types. Nine different movement… Show more

Dexterous finger movements can be decoded from neuronal action potentials acquired from a non-human primate using a chronically implanted Utah Electrode Array. We have developed an algorithm that can, after training, detect and classify individual and combined finger movements without any a priori knowledge of the data, task, or behavior. The algorithm is based on changes in the firing rates of individual neurons that are tuned for one or more finger movement types. Nine different movement types, which consisted of individual flexions, individual extensions, and combined flexions of the thumb, index finger, and middle finger, were decoded. The algorithm performed reliably on data recorded continuously during movement tasks, including a no-movement state, with an overall average sensitivity and specificity that were both >92%. These results demonstrate a viable algorithm for decoding dexterous finger movements under conditions similar to those required for a real-world neural prosthetic application. Show less

The ability to detect and classify individual and combined finger movements from neural data is rapidly advancing. The work that has been done has demonstrated the feasibility of decoding finger movements from acutely recorded neurons. There is a need for a recording model that meets the chronic requirements of a neuroprosthetic application and to address this need we have developed an algorithm that can detect and classify individual and combined finger movements using neuronal data acquired… Show more

The ability to detect and classify individual and combined finger movements from neural data is rapidly advancing. The work that has been done has demonstrated the feasibility of decoding finger movements from acutely recorded neurons. There is a need for a recording model that meets the chronic requirements of a neuroprosthetic application and to address this need we have developed an algorithm that can detect and classify individual and combined finger movements using neuronal data acquired from a chronically implanted Utah Electrode Array (UEA). The algorithm utilized the firing rates of individual neurons and performed with an average sensitivity and an average specificity that were both greater than 92% across all movement types. These results lend further support that a chronically implanted UEA is suitable for acquiring and decoding neuronal data and also demonstrate a decoding method that can detect and classify finger movements without any a priori knowledge of the data, task, or behavior. Show less