AI-Based Denoising of Head Impact Kinematics Measurements With Convolutional Neural Network for Traumatic Brain Injury Prediction.

OBJECTIVE: Wearable devices are developed to measure head impact kinematics but are intrinsically noisy because of the imperfect interface with human bodies. This study aimed to improve the head impact kinematics measurements obtained from instrumented mouthguards using deep learning to enhance traumatic brain injury (TBI) risk monitoring. METHODS: We developed one-dimensional convolutional neural network (1D-CNN) models to denoise mouthguard kinematics measurements for tri-axial linear acceleration and tri-axial angular velocity from 163 laboratory dummy head impacts. The performance of the denoising models was evaluated on three levels: kinematics, brain injury criteria, and tissue-level strain and strain rate. Additionally, we performed a blind test on an on-field dataset of 118 college football impacts and a test on 413 post-mortem human subject (PMHS) impacts. RESULTS: On the dummy head impacts, the denoised kinematics showed better correlation with reference kinematics, with relative reductions of 36% for pointwise root mean squared error and 56% for peak absolute error. Absolute errors in six brain injury criteria were reduced by a mean of 82%. For maximum principal strain and maximum principal strain rate, the mean error reduction was 35% and 69%, respectively. On the PMHS impacts, similar denoising effects were observed and the peak kinematics after denoising were more accurate (relative error reduction for 10% noisiest impacts was 75.6%). CONCLUSION: The 1D-CNN denoising models effectively reduced errors in mouthguard-derived kinematics measurements on dummy and PMHS impacts. SIGNIFICANCE: This study provides a novel approach for denoising head kinematics measurements in dummy and PMHS impacts, which can be further validated on more real-human kinematics data before real-world applications.

The role of occipital condyle and atlas anomalies on occipital cervical fusion outcomes in Chiari malformation type I with syringomyelia: a study from the Park-Reeves Syringomyelia Research Consortium.

OBJECTIVE: Congenital anomalies of the atlanto-occipital articulation may be present in patients with Chiari malformation type I (CM-I). However, it is unclear how these anomalies affect the biomechanical stability of the craniovertebral junction (CVJ) and whether they are associated with an increased incidence of occipitocervical fusion (OCF) following posterior fossa decompression (PFD). The objective of this study was to determine the prevalence of condylar hypoplasia and atlas anomalies in children with CM-I and syringomyelia. The authors also investigated the predictive contribution of these anomalies to the occurrence of OCF following PFD (PFD+OCF). METHODS: The authors analyzed the prevalence of condylar hypoplasia and atlas arch anomalies for patients in the Park-Reeves Syringomyelia Research Consortium database who underwent PFD+OCF. Condylar hypoplasia was defined by an atlanto-occipital joint axis angle (AOJAA) ≥ 130°. Atlas assimilation and arch anomalies were identified on presurgical radiographic imaging. This PFD+OCF cohort was compared with a control cohort of patients who underwent PFD alone. The control group was matched to the PFD+OCF cohort according to age, sex, and duration of symptoms at a 2:1 ratio. RESULTS: Clinical features and radiographic atlanto-occipital joint parameters were compared between 19 patients in the PFD+OCF cohort and 38 patients in the PFD-only cohort. Demographic data were not significantly different between cohorts (p > 0.05). The mean AOJAA was significantly higher in the PFD+OCF group than in the PFD group (144° ± 12° vs 127° ± 6°, p < 0.0001). In the PFD+OCF group, atlas assimilation and atlas arch anomalies were identified in 10 (53%) and 5 (26%) patients, respectively. These anomalies were absent (n = 0) in the PFD group (p < 0.001). Multivariate regression analysis identified the following 3 CVJ radiographic variables that were predictive of OCF occurrence after PFD: AOJAA ≥ 130° (p = 0.01), clivoaxial angle < 125° (p = 0.02), and occipital condyle-C2 sagittal vertical alignment (C-C2SVA) ≥ 5 mm (p = 0.01). A predictive model based on these 3 factors accurately predicted OCF following PFD (C-statistic 0.95). CONCLUSIONS: The authors' results indicate that the occipital condyle-atlas joint complex might affect the biomechanical integrity of the CVJ in children with CM-I and syringomyelia. They describe the role of the AOJAA metric as an independent predictive factor for occurrence of OCF following PFD. Preoperative identification of these skeletal abnormalities may be used to guide surgical planning and treatment of patients with complex CM-I and coexistent osseous pathology.

Neurosurgical Outcomes Among Non-English Speakers: A Systematic Review and a Framework for Future Research.

OBJECTIVE: In 2019, 22% of adults in the United States reported speaking a language other than English at home, representing 52% growth since 2000. This diversity in languages - and resulting possible communication barriers - represents a potential challenge to effective care. In this manuscript, we summarize clinical outcomes and healthcare utilization patterns of adult and pediatric neurosurgical patients who are non-English primary language speakers (NEPLS). METHODS: We systematically queried 5 databases from inception through October 2022. The Preferred Reporting Items for Systematic Reviews and Meta-Analyses guidelines were followed to identify studies for inclusion. The Newcastle-Ottawa Scale was used to assess the quality of studies. Additionally, a retrospective chart review was conducted to assess differences in postoperative communication patterns in a cohort of English and Spanish speaking patients with craniosynostosis at our institution. RESULTS: Our search yielded 442 abstracts; ten were included in the final cohort. Outcomes for 973 unique NEPLS with a neurosurgical condition were included; Spanish was the most represented language. Delivery and timing of surgical treatment was the most frequently reported metric; 75% of studies demonstrated a statistically significant delay in time to surgery or decreased likelihood for NEPLS to receive surgical treatment. Length of stay was reported in 3 studies; all demonstrated that NEPLS had longer length of stay. CONCLUSIONS: There is a paucity of literature reporting outcomes among NEPLS. It is critical to examine NEPLS patients' outcomes and experiences, as language barriers are potentially modifiable demographic factors. We present a framework that demonstrates opportunities for further research to improve quality of care.

Brain Deformation Estimation With Transfer Learning for Head Impact Datasets Across Impact Types.

OBJECTIVE: The machine-learning head model (MLHM) to accelerate the calculation of brain strain and strain rate, which are the predictors for traumatic brain injury (TBI), but the model accuracy was found to decrease sharply when the training/test datasets were from different head impacts types (i.e., car crash, college football), which limits the applicability of MLHMs to different types of head impacts and sports. Particularly, small sizes of target dataset for specific impact types with tens of impacts may not be enough to train an accurate impact-type-specific MLHM. METHODS: To overcome this, we propose data fusion and transfer learning to develop a series of MLHMs to predict the maximum principal strain (MPS) and maximum principal strain rate (MPSR). RESULTS: The strategies were tested on American football (338), mixed martial arts (457), reconstructed car crash (48) and reconstructed American football (36) and we found that the MLHMs developed with transfer learning are significantly more accurate in estimating MPS and MPSR than other models, with a mean absolute error (MAE) smaller than 0.03 in predicting MPS and smaller than [Formula: see text] in predicting MPSR on all target impact datasets. High performance in concussion detection was observed based on the MPS and MPSR estimated by the transfer-learning-based models. CONCLUSION: The MLHMs can be applied to various head impact types for rapidly and accurately calculating brain strain and strain rate. SIGNIFICANCE: This study enables developing MLHMs for the head impact type with limited availability of data, and will accelerate the applications of MLHMs.