Review of Machine Learning Algorithms for Diagnosing Mental Illness

OBJECTIVE: Enhanced technology in computer and internet has driven scale and quality of data to be improved in various areas including healthcare sectors. Machine Learning (ML) has played a pivotal role in efficiently analyzing those big data, but a general misunderstanding of ML algorithms still exists in applying them (e.g., ML techniques can settle a problem of small sample size, or deep learning is the ML algorithm). This paper reviewed the research of diagnosing mental illness using ML algorithm and suggests how ML techniques can be employed and worked in practice. METHODS: Researches about mental illness diagnostic using ML techniques were carefully reviewed. Five traditional ML algorithms-Support Vector Machines (SVM), Gradient Boosting Machine (GBM), Random Forest, Naïve Bayes, and K-Nearest Neighborhood (KNN)-frequently used for mental health area researches were systematically organized and summarized. RESULTS: Based on literature review, it turned out that Support Vector Machines (SVM), Gradient Boosting Machine (GBM), Random Forest, Naïve Bayes, and K-Nearest Neighborhood (KNN) were frequently employed in mental health area, but many researchers did not clarify the reason for using their ML algorithm though every ML algorithm has its own advantages. In addition, there were several studies to apply ML algorithms without fully understanding the data characteristics. CONCLUSION: Researchers using ML algorithms should be aware of the properties of their ML algorithms and the limitation of the results they obtained under restricted data conditions. This paper provides useful information of the properties and limitation of each ML algorithm in the practice of mental health.

Usefulness of Silicone Plate for Sellar Floor Reconstruction

No abstract available.

Electrically Evoked Stapedial Reflex in Cochlear Implantation / 대한이비인후과학회지

BACKGROUND AND OBJECTIVES: An essential part of the successful application of a cochlear implant is to adjust the stimulation levels to the patient's dynamic range. To test the subjective behavioral threshold (T-level) and the most comfortable loudness level (C-level), however, it may be troublesome or even impossible in very young children who are not able to perform such tasks appropriately. Therefore, utilization of objective measurements is important in children to estimate T- and C-levels. One of the objective measurements is the electrically evoked stapedial reflex (ESR) test. It has also been suggested that ESR may have applicability in confirming and quantifying electrical stimulation through a cochlear implant. The purpose of this study is to examine the feasibility of utilizing ESR to test device function during cochlear implant surgery and to predict ultimate behavioral T- & C-levels. MATERIALS AND METHODS: The ESR measurements were performed in 18 subjects during a cochlear implant surgery. Eight subjects received the Nucleus 22 multichannel cochlear implant (CI 22M) and 10 subjects received Nucleus 24 multichannel cochlear implant system (CI 24M). To test the ESRs, three electrodes, namely, apical electrode (E20), medial electrode (E12) and basal electrode (E5), were stimulated in each case. ESR thresholds were measured by visual observation of the stapedius muscle contraction in response to electrical stimulation delivered to the auditory nerve via a cochlear implant system. And the ESR thresholds were compared with the behavioral T- and C-levels in each electrode. RESULTS: ESRs were presented in 19/23 electrodes (82.6%) in subjects who received the CI 22M and 26/30 electrodes (86.7%) in subjects with CI 24M. The mean ESR threshold was the lowest in apical electrodes (E20) in both groups. In subjects with CI 22M, ESR thresholds were significantly correlated with behavioral T- and C-levels in apical electrodes (E20). In subjects with CI 24M, ESR thresholds were highly correlated with the C-levels in E5, E12, and E20. CONCLUSION: ESR provides an objective, accurate and rapid method to evaluate the device function during cochlear implant surgeries and in estimating T-/C- levels, which may be useful in the initial programming of younger children and the difficult-to-tests.

Changes of Ear Canal with Mandibular Movement / 대한이비인후과학회지

BACKGROUND AND OBJECTIVES: Although the necessity of hearing aids as one of the rehabilitation methods is increasing, many hearing impaired persons are avoiding the use of behind-the-ear (BTE) hearing aids and prefer the complete-in-the-canal (CIC) hearing aids. CIC hearing aids have cosmetic and acoustic advantages but have problems of acoustic feedback and discomfort in wearing. Changes in the external auditory canal (EAC) by mandibular movements are a primary factor for the acoustic feedback and discomfort. The purpose of this research is to study the pattern of changes in the EAC by mandibular movement using the Korean population as subjects and to get basic data concerning the manufacture of impressions and shells for CIC hearing aids. MATERIALS AND METHODS: Subjects were 19 men with a fit external ear diameter to make CIC hearing aid. Impressions were made by taking 5 different conditions of jaw into consideration : fully opened state, 1/2 opened state, biting state, chewing state and resting state. Impressions were scanned at intervals of O.1mm using a 3D Laser digitizing system. We measured anterior to posterior (AP) width and superior to inferior (SI) height of each impression at the first bend, interbend and the 2nd bend of 5 conditions. RESULTS: Compared with the resting state, statistically significant changes in the AP width of the ear canal were observed at the interbend in the following order : fully opened jaw, 1/2 opened jaw (p <0.05), at 1st bend: 1/2 opened jaw (p<0.05). The 2nd bend biting state showed maximal AP width of the ear canal (p<0.05). The SI height of ear canal revealed statistically insignificant changes. CONCLUSION: We recommend making of ear impressions in the chewing state and shells to be supported at the interbend of the ear canal. The SI dimension of a shell tip should be more ground than AP dimension.