A semi-supervised Support Vector Machine model for predicting the language outcomes following cochlear implantation based on pre-implant brain fMRI imaging

Artificial Intelligence in Otology and Neurotology

Clinical applications of artificial intelligence (AI) have grown exponentially with increasing computational power and Big Data. Data rich fields such as Otology and Neurotology are still in the infancy of harnessing the power of AI but are increasingly involved in training and developing ways to incorporate AI into patient care. Current studies involving AI are focused on accessible datasets; health care wearables, tabular data from electronic medical records, electrophysiologic measurements, imaging, and "omics" provide huge amounts of data to utilize. Health care wearables, such as hearing aids and cochlear implants, are a ripe environment for AI implementation.

Machine learning application in otology

This review presents a comprehensive history of Artificial Intelligence (AI) in the context of the revolutionary application of machine learning (ML) to medical research and clinical utilization, particularly for the benefit of researchers interested in the application of ML in otology. To this end, we discuss the key components of ML-input, output, and algorithms. In particular, some representation algorithms commonly used in medical research are discussed. Subsequently, we review ML applications in otology research, including diagnosis, influential identification, and surgical outcome prediction. In the context of surgical outcome prediction, specific surgical treatments, including cochlear implantation, active middle ear implantation, tympanoplasty, and vestibular schwannoma resection, are considered. Finally, we highlight the obstacles and challenges that need to be overcome in future research.

Prediction of Cochlear Implant Fitting by Machine Learning Techniques

OBJECTIVE

This study aimed to evaluate the differences in electrically evoked compound action potential (ECAP) thresholds and postoperative mapping current (T) levels between electrode types after cochlear implantation, the correlation between ECAP thresholds and T levels, and the performance of machine learning techniques in predicting postoperative T levels.

STUDY DESIGN

Retrospective case review.

SETTING

Tertiary hospital.

PATIENTS

We reviewed the charts of 124 ears of children with severe-to-profound hearing loss who had undergone cochlear implantation.

INTERVENTIONS

We compared ECAP thresholds and T levels from different electrodes, calculated correlations between ECAP thresholds and T levels, and created five prediction models of T levels at switch-on and 6 months after surgery.

MAIN OUTCOME MEASURES

The accuracy of prediction in postoperative mapping current (T) levels.

RESULTS

The ECAP thresholds of the slim modiolar electrodes were significantly lower than those of the straight electrodes on the apical side. However, there was no significant difference in the neural response telemetry thresholds between the two electrodes on the basal side. Lasso regression achieved the most accurate prediction of T levels at switch-on, and the random forest algorithm achieved the most accurate prediction of T levels 6 months after surgery in this dataset.

CONCLUSION

Machine learning techniques could be useful for accurately predicting postoperative T levels after cochlear implantation in children.

Machine learning-based prediction of the outcomes of cochlear implantation in patients with inner ear malformation

OBJECTIVE

The objectives of this study are twofold: first, to visualize the structure of malformed cochleae through image reconstruction; and second, to develop a predictive model for postoperative outcomes of cochlear implantation (CI) in patients diagnosed with cochlear hypoplasia (CH) and incomplete partition (IP) malformation.

METHODS

The clinical data from patients diagnosed with cochlear hypoplasia (CH) and incomplete partition (IP) malformation who underwent cochlear implantation (CI) at Beijing Tongren Hospital between January 2016 and August 2020 were collected. Radiological features were analyzed through 3D segmentation of the cochlea. Postoperative auditory speech rehabilitation outcomes were evaluated using the Categories of Auditory Performance (CAP) and the Speech Intelligibility Rating (SIR). This study aimed to investigate the relationship between cochlear parameters and postoperative outcomes. Additionally, a predictive model for postoperative outcomes was developed using the K-nearest neighbors (KNN) algorithm.

RESULTS

In our study, we conducted feature selection by using patients' imaging and audiological attributes. This process involved methods such as the removal of missing values, correlation analysis, and chi-square tests. The findings indicated that two specific features, cochlear volume (V) and cochlear canal length (CDL), significantly contributed to predicting the outcomes of hearing and speech rehabilitation for patients with inner ear malformations. In terms of hearing rehabilitation, the KNN classification achieved an accuracy of 93.3%. Likewise, for speech rehabilitation, the KNN classification demonstrated an accuracy of 86.7%.

CONCLUSION

The measurements obtained from the 3D reconstruction model hold significant clinical relevance. Despite the considerable variability in cochlear morphology across individuals, radiological features remain effective in predicting cochlear implantation (CI) prognosis for patients with inner ear malformations. The utilization of 3D segmentation techniques and the developed predictive model can assist surgeons in conducting preoperative cochlear structural measurements for patients with inner ear malformations. This, in turn, can offer a more informed perspective on the anticipated outcomes of cochlear implantation.

Recent advances in Otology: Current landscape and future direction

Hearing is an essential sensation, and its deterioration leads to a significant decrease in the quality of life. Thus, great efforts have been made by otologists to preserve and recover hearing. Our knowledge regarding the field of otology has progressed with advances in technology, and otologists have sought to develop novel approaches in the field of otologic surgery to achieve higher hearing recovery or preservation rates. This requires knowledge regarding the anatomy of the temporal bone and the physiology of hearing. Basic research in the field of otology has progressed with advances in molecular biology and genetics. This review summarizes the current views and recent advances in the field of otology and otologic surgery, especially from the viewpoint of young Japanese clinician-scientists, and presents the perspectives and future directions for several topics in the field of otology. This review will aid next-generation researchers in understanding the recent advances and future challenges in the field of otology.

Predicting synkinesis caused by Bell's palsy or Ramsay Hunt syndrome using machine learning-based logistic regression

Objective

To investigate whether machine learning (ML)-based algorithms, namely logistic regression (LR), random forest (RF), k-nearest neighbor (k-NN), and gradient-boosting decision tree (GBDT), utilizing early post-onset parameters can predict facial synkinesis resulting from Bell's palsy or Ramsay Hunt syndrome more accurately than the conventional statistics-based LR.

Methods

This retrospective study included 362 patients who presented to a facial palsy outpatient clinic. Median follow-up of synkinesis-positive and -negative patients was 388 (range, 177-1922) and 198 (range, 190-3021) days, respectively. Electrophysiological examinations were performed, and the rate of synkinesis in Bell's palsy and Ramsay Hunt syndrome was evaluated. Sensitivity and specificity were assessed using statistics-based LR; and electroneurography (ENoG) value, the difference in the nerve excitability test (NET), and scores of the subjective Yanagihara scaling system were evaluated using early post-onset parameters with ML-based LR, RF, k-NN, and GBDT.

Results

Synkinesis rate in Bell's palsy and Ramsay Hunt syndrome was 20.2% (53/262) and 40.0% (40/100), respectively. Sensitivity and specificity obtained with statistics-based LR were 0.796 and 0.806, respectively, and the area under the receiver operating characteristic curve (AUC) was 0.87. AUCs measured using ML-based LR of "ENoG," "difference in NET," "Yanagihara," and all three components ("all") were 0.910, 0.834, 0.711, and 0.901, respectively.

Conclusion

ML-based LR model shows potential in predicting facial synkinesis probability resulting from Bell's palsy or Ramsay Hunt syndrome and has comparable reliability to the conventional statistics-based LR.

Level of Evidence

3.

Application of Machine Learning to Predict Hearing Outcomes of Tympanoplasty

OBJECTIVE

This retrospective study aimed to evaluate the performance of machine learning techniques in predicting air-bone gap after tympanoplasty compared with conventional scoring models and to identify the influential factors.

METHODS

We reviewed the charts of 105 patients (114 ears) with chronic otitis media who underwent tympanoplasty. Two numerical scoring systems (middle ear risk index [MERI] and ossiculoplasty outcome parameter staging [OOPS]) and three algorithms (random forest [RF], support vector machine [SVM], and k nearest neighbor [kNN]) were created. Experimental variables included age, preoperative air-bone gap, soft-tissue density lesion in the tympanic cavity in CT, otorrhea, surgical history, ossicular bone problems in CT, tympanic perforation location, perforation type (central or marginal), grafting material, smoking history, endoscopy use, and the operator whose experience was 20 years or longer, or shorter. Binary classification, postoperative air-bone gap ≤15 or >15 dB, and multiclass classification, classification into seven categories by 10 dB, were performed, and the percentages of correct prediction were calculated. The importance of features in the RF model was calculated to identify influential factors.

RESULTS

The percentages of correct prediction in binary classification were 62.3%, 72.8%, 81.5%, 81.5%, and 81.5% in MERI, OOPS, RF, SVM, and kNN, respectively, and those in multiclass classification were 29.8%, 21.9%, 63.1%, 44.7%, and 50% in the same order. The RF model suggested larger preoperative air-bone gap, and older age could make the postoperative air-bone gap larger.

CONCLUSION

The machine learning techniques, especially the RF model, are promising methods for precise postoperative air-bone gap prediction.

LEVEL OF EVIDENCE

4 Laryngoscope, 133:2371-2378, 2023.

Machine Learning and Prediction in Fetal, Infant, and Toddler Neuroimaging: A Review and Primer

Predictive models in neuroimaging are increasingly designed with the intent to improve risk stratification and support interventional efforts in psychiatry. Many of these models have been developed in samples of children school-aged or older. Nevertheless, despite growing evidence that altered brain maturation during the fetal, infant, and toddler (FIT) period modulates risk for poor mental health outcomes in childhood, these models are rarely implemented in FIT samples. Applications of predictive modeling in children of these ages provide an opportunity to develop powerful tools for improved characterization of the neural mechanisms underlying development. To facilitate the broader use of predictive models in FIT neuroimaging, we present a brief primer and systematic review on the methods used in current predictive modeling FIT studies. Reflecting on current practices in more than 100 studies conducted over the past decade, we provide an overview of topics, modalities, and methods commonly used in the field and under-researched areas. We then outline ethical and future considerations for neuroimaging researchers interested in predicting health outcomes in early life, including researchers who may be relatively new to either advanced machine learning methods or using FIT data. Altogether, the last decade of FIT research in machine learning has provided a foundation for accelerating the prediction of early-life trajectories across the full spectrum of illness and health.

Machine Learning-Based Prediction of the Outcomes of Cochlear Implantation in Patients With Cochlear Nerve Deficiency and Normal Cochlea: A 2-Year Follow-Up of 70 Children

Cochlear nerve deficiency (CND) is often associated with variable outcomes of cochlear implantation (CI). We assessed previous investigations aiming to identify the main factors that determine CI outcomes, which would enable us to develop predictive models. Seventy patients with CND and normal cochlea who underwent CI surgery were retrospectively examined. First, using a data-driven approach, we collected demographic information, radiographic measurements, audiological findings, and audition and speech assessments. Next, CI outcomes were evaluated based on the scores obtained after 2 years of CI from the Categories of Auditory Performance index, Speech Intelligibility Rating, Infant/Toddler Meaningful Auditory Integration Scale or Meaningful Auditory Integration Scale, and Meaningful Use of Speech Scale. Then, we measured and averaged the audiological and radiographic characteristics of the patients to form feature vectors, adopting a multivariate feature selection method, called stability selection, to select the features that were consistent within a certain range of model parameters. Stability selection analysis identified two out of six characteristics, namely the vestibulocochlear nerve (VCN) area and the number of nerve bundles, which played an important role in predicting the hearing and speech rehabilitation results of CND patients. Finally, we used a parameter-optimized support vector machine (SVM) as a classifier to study the postoperative hearing and speech rehabilitation of the patients. For hearing rehabilitation, the accuracy rate was 71% for both the SVM classification and the area under the curve (AUC), whereas for speech rehabilitation, the accuracy rate for SVM classification and AUC was 93% and 94%, respectively. Our results identified that a greater number of nerve bundles and a larger VCN area were associated with better CI outcomes. The number of nerve bundles and VCN area can predict CI outcomes in patients with CND. These findings can help surgeons in selecting the side for CI and provide reasonable expectations for the outcomes of CI surgery.

Functional MRI research involving healthy children: Ethics, safety and recommended procedures

AIM

This specific review aims to expose clinicians, researchers and administrators in hospitals to the importance, procedures and safety of fMRI studies to promote the increased utilisation of such studies in different geographical places worldwide. The child's brain is developing rapidly, both structurally and functionally. These functional changes can only be detected using functional scans generated from an MRI machine and referred to as a functional MRI (fMRI). This method may be used clinically in complex medical and surgical conditions (e.g., epilepsy surgery), but these days are often used for research purposes. However, due to ethical and logistical considerations, fMRI in the paediatric population is not widely and equally used in different geographical places.

CONCLUSIONS

The benefits of using this method to define the functional changes occurring in the developing brain are discussed in this review, along with desensitisation methods recommended when working with this vulnerable population in research and even in a clinical setting.

Brain Morphological Modifications in Congenital and Acquired Auditory Deprivation: A Systematic Review and Coordinate-Based Meta-Analysis

Neuroplasticity following deafness has been widely demonstrated in both humans and animals, but the anatomical substrate of these changes is not yet clear in human brain. However, it is of high importance since hearing loss is a growing problem due to aging population. Moreover, knowing these brain changes could help to understand some disappointing results with cochlear implant, and therefore could improve hearing rehabilitation. A systematic review and a coordinate-based meta-analysis were realized about the morphological brain changes highlighted by MRI in severe to profound hearing loss, congenital and acquired before or after language onset. 25 papers were included in our review, concerning more than 400 deaf subjects, most of them presenting prelingual deafness. The most consistent finding is a volumetric decrease in gray matter around bilateral auditory cortex. This change was confirmed by the coordinate-based meta-analysis which shows three converging clusters in this region. The visual areas of deaf children is also significantly impacted, with a decrease of the volume of both gray and white matters. Finally, deafness is responsible of a gray matter increase within the cerebellum, especially at the right side. These results are largely discussed and compared with those from deaf animal models and blind humans, which demonstrate for example a much more consistent gray matter decrease along their respective primary sensory pathway. In human deafness, a lot of other factors than deafness could interact on the brain plasticity. One of the most important is the use of sign language and its age of acquisition, which induce among others changes within the hand motor region and the visual cortex. But other confounding factors exist which have been too little considered in the current literature, such as the etiology of the hearing impairment, the speech-reading ability, the hearing aid use, the frequent associated vestibular dysfunction or neurocognitive impairment. Another important weakness highlighted by this review concern the lack of papers about postlingual deafness, whereas it represents most of the deaf population. Further studies are needed to better understand these issues, and finally try to improve deafness rehabilitation.

Using Clinical Audiologic Measures to Determine Cochlear Implant Candidacy

INTRODUCTION

Only a small percentage (6-10%) of patients who are candidates receive cochlear implants (CIs). One potential reason contributing to low usage rates may be confusion regarding which patients to refer for CI evaluation. The extent to which information provided by standard clinical audiologic assessments is sufficient for selecting appropriate CI evaluation referrals is uncertain. The objective of this study is to evaluate the capacity of standard clinical audiologic measures to differentiate CI candidates from noncandidates.

METHOD

The study design is a retrospective review of a prospectively maintained CI database from a university-based tertiary medical center of 518 patients undergoing CI evaluations from 2012 to 2020. Each ear of each patient was treated as an independent value. Receiver operating characteristic (ROCs) curves were constructed using aided AzBio sentence recognition scores in quiet and aided AzBio +10 dB signal-to-noise ratio scores <60% as binary classifiers for CI candidacy. For each ROC, we examined the capacity of multiple pure-tone thresholds, pure-tone average (PTA), and CNC word recognition scores (WRSs) measured under earphones to determine CI candidacy. Area under the curve ROC (AUC-ROC) values were calculated to demonstrate the capacity of each model to differentiate CI candidates from noncandidates.

RESULTS

Variables with the greatest capacity to accurately differentiate CI candidates from noncandidates using aided AzBio in quiet scores were earphone CNC WRS, earphone pure-tone threshold at 1,000 Hz, and earphone PTA (AUC-ROC values = 0.86-0.88). Using aided AzBio +10 scores as the measure for candidacy, only CNC word recognition had a fair capacity to identify candidates (AUC-ROC value = 0.73). Based on the ROCs, a 1,000 Hz pure-tone threshold >50 dB HL, PTA >57 dB HL, and a monosyllabic WRS <60% can each serve as individual indicators for referral for CI evaluations.

CONCLUSION

The current study provides initial indicators for referral and a first step at developing evidence-based criteria for CI evaluation referral using standard audiologic assessments.

Machine Learning Technique Reveals Prognostic Factors of Vibrant Soundbridge for Conductive or Mixed Hearing Loss Patients

OBJECTIVES

Vibrant Soundbridge (VSB) was developed for treatment of hearing loss, but clinical outcomes vary and prognostic factors predicting the success of the treatment remain unknown. We examined clinical outcomes of VSB for conductive or mixed hearing loss, prognostic factors by analyzing prediction models, and cut-off values to predict the outcomes.

STUDY DESIGN

Retrospective chart review.

SETTING

Tertiary care hospital.

PATIENTS

Thirty patients who underwent VSB surgery from January 2017 to December 2019 at our hospital.

INTERVENTION

Audiological tests were performed prior to and 3 months after surgery; patients completed questionnaires 3 months after surgery.

MAIN OUTCOME MEASURES

We used a multiregression and the random forest algorithm for predictions. Mean absolute errors and coefficient of determinations were calculated to estimate prediction accuracies. Coefficient values in the multiregression model and the importance of features in the random forest model were calculated to clarify prognostic factors. Receiver operation characteristic curves were plotted.

RESULTS

All audiological outcomes improved after surgery. The random forest model (mean absolute error: 0.06) recorded more accuracy than the multiregression model (mean absolute error: 0.12). Speech discrimination score in a silent context in patients with hearing aids was the most influential factor (coefficient value: 0.51, featured value: 0.71). The candidate cut-off value was 36% (sensitivity: 89%, specificity: 75%).

CONCLUSIONS

VSB is an effective treatment for conductive or mixed hearing loss. Machine learning demonstrated more precise predictions, and speech discrimination scores in a silent context in patients with hearing aids were the most important factor in predicting clinical outcomes.

Recent Advances in the Application of Artificial Intelligence in Otorhinolaryngology-Head and Neck Surgery

This study presents an up-to-date survey of the use of artificial intelligence (AI) in the field of otorhinolaryngology, considering opportunities, research challenges, and research directions. We searched PubMed, the Cochrane Central Register of Controlled Trials, Embase, and the Web of Science. We initially retrieved 458 articles. The exclusion of non-English publications and duplicates yielded a total of 90 remaining studies. These 90 studies were divided into those analyzing medical images, voice, medical devices, and clinical diagnoses and treatments. Most studies (42.2%, 38/90) used AI for image-based analysis, followed by clinical diagnoses and treatments (24 studies). Each of the remaining two subcategories included 14 studies. Machine learning and deep learning have been extensively applied in the field of otorhinolaryngology. However, the performance of AI models varies and research challenges remain.

Prediction of the Functional Status of the Cochlear Nerve in Individual Cochlear Implant Users Using Machine Learning and Electrophysiological Measures

OBJECTIVES

This study aimed to create an objective predictive model for assessing the functional status of the cochlear nerve (CN) in individual cochlear implant (CI) users.

DESIGN

Study participants included 23 children with cochlear nerve deficiency (CND), 29 children with normal-sized CNs (NSCNs), and 20 adults with various etiologies of hearing loss. Eight participants were bilateral CI users and were tested in both ears. As a result, a total of 80 ears were tested in this study. All participants used Cochlear Nucleus CIs in their test ears. For each participant, the CN refractory recovery function and input/output (I/O) function were measured using electrophysiological measures of the electrically evoked compound action potential (eCAP) at three electrode sites across the electrode array. Refractory recovery time constants were estimated using statistical modeling with an exponential decay function. Slopes of I/O functions were estimated using linear regression. The eCAP parameters used as input variables in the predictive model were absolute refractory recovery time estimated based on the refractory recovery function, eCAP threshold, slope of the eCAP I/O function, and negative-peak (i.e., N1) latency. The output variable of the predictive model was CN index, an indicator for the functional status of the CN. Predictive models were created by performing linear regression, support vector machine regression, and logistic regression with eCAP parameters from children with CND and the children with NSCNs. One-way analysis of variance with post hoc analysis with Tukey's honest significant difference criterion was used to compare study variables among study groups.

RESULTS

All three machine learning algorithms created two distinct distributions of CN indices for children with CND and children with NSCNs. Variations in CN index when calculated using different machine learning techniques were observed for adult CI users. Regardless of these variations, CN indices calculated using all three techniques in adult CI users were significantly correlated with Consonant-Nucleus-Consonant word and AzBio sentence scores measured in quiet. The five oldest CI users had smaller CN indices than the five youngest CI users in this study.

CONCLUSIONS

The functional status of the CN for individual CI users was estimated by our newly developed analytical models. Model predictions of CN function for individual adult CI users were positively and significantly correlated with speech perception performance. The models presented in this study may be useful for understanding and/or predicting CI outcomes for individual patients.

Machine Learning and Cochlear Implantation-A Structured Review of Opportunities and Challenges

OBJECTIVE

The use of machine learning technology to automate intellectual processes and boost clinical process efficiency in medicine has exploded in the past 5 years. Machine learning excels in automating pattern recognition and in adapting learned representations to new settings. Moreover, machine learning techniques have the advantage of incorporating complexity and are free from many of the limitations of traditional deterministic approaches. Cochlear implants (CI) are a unique fit for machine learning techniques given the need for optimization of signal processing to fit complex environmental scenarios and individual patients' CI MAPping. However, there are many other opportunities where machine learning may assist in CI beyond signal processing. The objective of this review was to synthesize past applications of machine learning technologies for pediatric and adult CI and describe novel opportunities for research and development.

DATA SOURCES

The PubMed/MEDLINE, EMBASE, Scopus, and ISI Web of Knowledge databases were mined using a directed search strategy to identify the nexus between CI and artificial intelligence/machine learning literature.

STUDY SELECTION

Non-English language articles, articles without an available abstract or full-text, and nonrelevant articles were manually appraised and excluded. Included articles were evaluated for specific machine learning methodologies, content, and application success.

DATA SYNTHESIS

The database search identified 298 articles. Two hundred fifty-nine articles (86.9%) were excluded based on the available abstract/full-text, language, and relevance. The remaining 39 articles were included in the review analysis. There was a marked increase in year-over-year publications from 2013 to 2018. Applications of machine learning technologies involved speech/signal processing optimization (17; 43.6% of articles), automated evoked potential measurement (6; 15.4%), postoperative performance/efficacy prediction (5; 12.8%), and surgical anatomy location prediction (3; 7.7%), and 2 (5.1%) in each of robotics, electrode placement performance, and biomaterials performance.

CONCLUSION

The relationship between CI and artificial intelligence is strengthening with a recent increase in publications reporting successful applications. Considerable effort has been directed toward augmenting signal processing and automating postoperative MAPping using machine learning algorithms. Other promising applications include augmenting CI surgery mechanics and personalized medicine approaches for boosting CI patient performance. Future opportunities include addressing scalability and the research and clinical communities' acceptance of machine learning algorithms as effective techniques.

Iterative processes: a review of semi-supervised machine learning in rehabilitation science

Purpose: To define semi-supervised machine learning (SSML) and explore current and potential applications of this analytic strategy in rehabilitation research.Method: We conducted a scoping review using PubMed, GoogleScholar and Medline. Studies were included if they: (1) described a semi-supervised approach to apply machine learning algorithms during data analysis and (2) examined constructs encompassed by the International Classification of Functioning, Disability and Health (ICF). The first two authors reviewed identified articles and recorded study and participant characteristics. The ICF domain used in each study was also identified.Results: After combining information from the eight studies, we established that SSML was a feasible approach for analysis of complex data in rehabilitation research. We also determined that semi-supervised approaches may be more accurate than supervised machine learning approaches.Conclusions: A semi-supervised approach to machine learning has potential to enhance our understanding of complex data sets in rehabilitation science. SSML mirrors the iterative process of rehabilitation, making this approach ideal for calibrating devices, classifying activities or identifying just-in-time interventions. Rehabilitation scientists who are interested in conducting SSML should collaborate with data scientists to advance the application of this approach within our field.Implications for rehabilitationSemi-supervised machine learning applications may be a feasible approach for analyses of complex data sets in rehabilitation research.Semi-supervised machine learning approaches uses a combination of labelled and unlabelled data to produce accurate predictive models, thereby requiring less user-input data than other machine learning approaches (i.e., supervised, unsupervised), reducing resource cost and user-burden.Semi-supervised machine learning is an iterative process that, when applied to rehabilitation assessment and outcomes, could produce accurate personalized models for treatment.Rehabilitation researchers and data scientists should collaborate to implement semi-supervised machine learning approaches in rehabilitation research, optimizing the power of large datasets that are becoming more readily available within the field (e.g., EEG signals, sensors, smarthomes).

Impedance Measures During in vitro Cochlear Implantation Predict Array Positioning

OBJECTIVE

Improper electrode placement during cochlear implant (CI) insertion can adversely affect speech perception outcomes. However, the intraoperative methods to determine positioning are limited. Because measures of electrode impedance can be made quickly, the goal of this study was to assess the relationship between CI impedance and proximity to adjacent structures.

METHODS

An Advanced Bionics CI array was inserted into a clear, plastic cochlea one electrode contact at a time in a saline bath (nine trials). At each insertion depth, response to biphasic current pulses was used to calculate access resistance (Ra), polarization resistance (Rp), and polarization capacitance (Cp). These measures were correlated to actual proximity as assessed by microscopy using linear regression models.

RESULTS

Impedance increased with insertion depth and proximity to the inner wall. Specifically, Ra increased, Cp decreased, and Rp slightly increased. Incorporating all impedance measures afforded a prediction model (r = 0.88) while optimizing for sub-mm positioning afforded a model with 78.3% specificity.

CONCLUSION

Impedance in vitro greatly changes with electrode insertion depth and proximity to adjacent structures in a predicable manner.

SIGNIFICANCE

Assessing proximity of the CI to adjacent structures is a significant first step in qualifying the electrode-neural interface. This information should aid in CI fitting, which should help maximize hearing and speech outcomes with a CI. Additionally, knowledge of the relationship between impedance and positioning could have utility in other tissue implants in the brain, retina, or spinal cord.

Use of machine-learning algorithms to determine features of systolic blood pressure variability that predict poor outcomes in hypertensive patients

Background

We re-analyzed data from the Systolic Blood Pressure Intervention Trial (SPRINT) trial to identify features of systolic blood pressure (SBP) variability that portend poor cardiovascular outcomes using a nonlinear machine-learning algorithm.

Methods

We included all patients who completed 1 year of the study without reaching any primary endpoint during the first year, specifically: myocardial infarction, other acute coronary syndromes, stroke, heart failure or death from a cardiovascular event (n = 8799; 94%). In addition to clinical variables, features representing longitudinal SBP trends and variability were determined and combined in a random forest algorithm, optimized using cross-validation, using 70% of patients in the training set. Area under the curve (AUC) was measured using a 30% testing set. Finally, feature importance was determined by minimizing node impurity averaging over all trees in the forest for a specific feature.

Results

A total of 365 patients (4.1%) reached the combined primary outcome over 37 months of follow-up. The random forest classifier had an AUC of 0.71 on the testing set. The 10 most significant features selected in order of importance by the automated algorithm included the urine albumin/creatinine (CR) ratio, estimated glomerular filtration rate, age, serum CR, history of subclinical cardiovascular disease (CVD), cholesterol, a variable representing SBP signals using wavelet transformation, high-density lipoprotein, the 90th percentile of SBP and triglyceride level.

Conclusions

We successfully demonstrated use of random forest algorithm to define best prognostic longitudinal SBP representations. In addition to known risk factors for CVD, transformed variables for time series SBP measurements were found to be important in predicting poor cardiovascular outcomes and require further evaluation.

Decision Variants for the Automatic Determination of Optimal Feature Subset in RF-RFE

Feature selection, which identifies a set of most informative features from the original feature space, has been widely used to simplify the predictor. Recursive feature elimination (RFE), as one of the most popular feature selection approaches, is effective in data dimension reduction and efficiency increase. A ranking of features, as well as candidate subsets with the corresponding accuracy, is produced through RFE. The subset with highest accuracy (HA) or a preset number of features (PreNum) are often used as the final subset. However, this may lead to a large number of features being selected, or if there is no prior knowledge about this preset number, it is often ambiguous and subjective regarding final subset selection. A proper decision variant is in high demand to automatically determine the optimal subset. In this study, we conduct pioneering work to explore the decision variant after obtaining a list of candidate subsets from RFE. We provide a detailed analysis and comparison of several decision variants to automatically select the optimal feature subset. Random forest (RF)-recursive feature elimination (RF-RFE) algorithm and a voting strategy are introduced. We validated the variants on two totally different molecular biology datasets, one for a toxicogenomic study and the other one for protein sequence analysis. The study provides an automated way to determine the optimal feature subset when using RF-RFE.

Neural preservation underlies speech improvement from auditory deprivation in young cochlear implant recipients

Although cochlear implantation enables some children to attain age-appropriate speech and language development, communicative delays persist in others, and outcomes are quite variable and difficult to predict, even for children implanted early in life. To understand the neurobiological basis of this variability, we used presurgical neural morphological data obtained from MRI of individual pediatric cochlear implant (CI) candidates implanted younger than 3.5 years to predict variability of their speech-perception improvement after surgery. We first compared neuroanatomical density and spatial pattern similarity of CI candidates to that of age-matched children with normal hearing, which allowed us to detail neuroanatomical networks that were either affected or unaffected by auditory deprivation. This information enables us to build machine-learning models to predict the individual children's speech development following CI. We found that regions of the brain that were unaffected by auditory deprivation, in particular the auditory association and cognitive brain regions, produced the highest accuracy, specificity, and sensitivity in patient classification and the most precise prediction results. These findings suggest that brain areas unaffected by auditory deprivation are critical to developing closer to typical speech outcomes. Moreover, the findings suggest that determination of the type of neural reorganization caused by auditory deprivation before implantation is valuable for predicting post-CI language outcomes for young children.

A Computational Model for the Automatic Diagnosis of Attention Deficit Hyperactivity Disorder Based on Functional Brain Volume

In this paper, we investigated the problem of computer-aided diagnosis of Attention Deficit Hyperactivity Disorder (ADHD) using machine learning techniques. With the ADHD-200 dataset, we developed a Support Vector Machine (SVM) model to classify ADHD patients from typically developing controls (TDCs), using the regional brain volumes as predictors. Conventionally, the volume of a brain region was considered to be an anatomical feature and quantified using structural magnetic resonance images. One major contribution of the present study was that we had initially proposed to measure the regional brain volumes using fMRI images. Brain volumes measured from fMRI images were denoted as functional volumes, which quantified the volumes of brain regions that were actually functioning during fMRI imaging. We compared the predictive power of functional volumes with that of regional brain volumes measured from anatomical images, which were denoted as anatomical volumes. The former demonstrated higher discriminative power than the latter for the classification of ADHD patients vs. TDCs. Combined with our two-step feature selection approach which integrated prior knowledge with the recursive feature elimination (RFE) algorithm, our SVM classification model combining functional volumes and demographic characteristics achieved a balanced accuracy of 67.7%, which was 16.1% higher than that of a relevant model published previously in the work of Sato et al. Furthermore, our classifier highlighted 10 brain regions that were most discriminative in distinguishing between ADHD patients and TDCs. These 10 regions were mainly located in occipital lobe, cerebellum posterior lobe, parietal lobe, frontal lobe, and temporal lobe. Our present study using functional images will likely provide new perspectives about the brain regions affected by ADHD.

Wavelet Entropy and Directed Acyclic Graph Support Vector Machine for Detection of Patients with Unilateral Hearing Loss in MRI Scanning

Highlights

We develop computer-aided diagnosis system for unilateral hearing loss detection in structural magnetic resonance imaging.

Wavelet entropy is introduced to extract image global features from brain images. Directed acyclic graph is employed to endow support vector machine an ability to handle multi-class problems.

The developed computer-aided diagnosis system achieves an overall accuracy of 95.1% for this three-class problem of differentiating left-sided and right-sided hearing loss from healthy controls.

Aim: Sensorineural hearing loss (SNHL) is correlated to many neurodegenerative disease. Now more and more computer vision based methods are using to detect it in an automatic way.

Materials: We have in total 49 subjects, scanned by 3.0T MRI (Siemens Medical Solutions, Erlangen, Germany). The subjects contain 14 patients with right-sided hearing loss (RHL), 15 patients with left-sided hearing loss (LHL), and 20 healthy controls (HC).

Method: We treat this as a three-class classification problem: RHL, LHL, and HC. Wavelet entropy (WE) was selected from the magnetic resonance images of each subjects, and then submitted to a directed acyclic graph support vector machine (DAG-SVM).

Results: The 10 repetition results of 10-fold cross validation shows 3-level decomposition will yield an overall accuracy of 95.10% for this three-class classification problem, higher than feedforward neural network, decision tree, and naive Bayesian classifier.

Conclusions: This computer-aided diagnosis system is promising. We hope this study can attract more computer vision method for detecting hearing loss.

A semi-supervised Support Vector Machine model for predicting the language outcomes following cochlear implantation based on pre-implant brain fMRI imaging

INTRODUCTION

We developed a machine learning model to predict whether or not a cochlear implant (CI) candidate will develop effective language skills within 2 years after the CI surgery by using the pre-implant brain fMRI data from the candidate.

METHODS

The language performance was measured 2 years after the CI surgery by the Clinical Evaluation of Language Fundamentals-Preschool, Second Edition (CELF-P2). Based on the CELF-P2 scores, the CI recipients were designated as either effective or ineffective CI users. For feature extraction from the fMRI data, we constructed contrast maps using the general linear model, and then utilized the Bag-of-Words (BoW) approach that we previously published to convert the contrast maps into feature vectors. We trained both supervised models and semi-supervised models to classify CI users as effective or ineffective.

RESULTS

Compared with the conventional feature extraction approach, which used each single voxel as a feature, our BoW approach gave rise to much better performance for the classification of effective versus ineffective CI users. The semi-supervised model with the feature set extracted by the BoW approach from the contrast of speech versus silence achieved a leave-one-out cross-validation AUC as high as 0.97. Recursive feature elimination unexpectedly revealed that two features were sufficient to provide highly accurate classification of effective versus ineffective CI users based on our current dataset.

CONCLUSION

We have validated the hypothesis that pre-implant cortical activation patterns revealed by fMRI during infancy correlate with language performance 2 years after cochlear implantation. The two brain regions highlighted by our classifier are potential biomarkers for the prediction of CI outcomes. Our study also demonstrated the superiority of the semi-supervised model over the supervised model. It is always worthwhile to try a semi-supervised model when unlabeled data are available.