A Study on Voice Measures in Patients with Parkinson's Disease.

PURPOSE: This research aims to identify acoustic features which can distinguish patients with Parkinson's disease (PD patients) and healthy speakers. METHODS: Thirty PD patients and 30 healthy speakers were recruited in the experiment, and their speech was collected, including three vowels (/i/, /a/, and /u/) and nine consonants (/p/, /pÊ°/, /t/, /tÊ°/, /k/, /kÊ°/, /l/, /m/, and /n/). Acoustic features like fundamental frequency (F0), Jitter, Shimmer, harmonics-to-noise ratio (HNR), first formant (F1), second formant (F2), third formant (F3), first bandwidth (B1), second bandwidth (B2), third bandwidth (B3), voice onset, voice onset time were analyzed in our experiment. Two-sample independent t test and the nonparametric Mann-Whitney U (MWU) test were carried out alternatively to compare the acoustic measures between the PD patients and healthy speakers. In addition, after figuring out the effective acoustic features for distinguishing PD patients and healthy speakers, we adopted two methods to detect PD patients: (1) Built classifiers based on the effective acoustic features and (2) Trained support vector machine classifiers via the effective acoustic features. RESULTS: Significant differences were found between the male PD group and the male health control in vowel /i/ (Jitter and Shimmer) and /a/ (Shimmer and HNR). Among female subjects, significant differences were observed in F0 standard deviation (F0 SD) of /u/ between the two groups. Additionally, significant differences between PD group and health control were also found in the F3 of /i/ and /n/, whereas other acoustic features showed no significant differences between the two groups. The HNR of vowel /a/ performed the best classification accuracy compared with the other six acoustic features above found to distinguish PD patients and healthy speakers. CONCLUSIONS: PD can cause changes in the articulation and phonation of PD patients, wherein increases or decreases occur in some acoustic features. Therefore, the use of acoustic features to detect PD is expected to be a low-cost and large-scale diagnostic method.

Carotid Ultrasound Combined with Plasma Atherosclerosis Index (AIP) in the Severe Analysis of Coronary Heart Disease Treated with Atorvastatin.

Background: Atherosclerotic coronary heart disease (CHD) stands as a paramount cardiovascular concern and the primary cause of mortality. To underscore the significance of our study, it is crucial to highlight the existing gaps in current diagnostic methods and prognostic assessments of CHD. By addressing these gaps, our research aims to contribute valuable insights and advancements in the understanding and management of this prevalent cardiovascular condition. Objective: The primary objective of this study is to investigate the correlation between carotid ultrasound, the Atherogenic Index of Plasma (AIP), and the severity of CHD. Methods: We enrolled 59 patients diagnosed with coronary heart disease and categorized them into two groups (multi-vessel and single-vessel disease groups) based on disease severity. The study employed carotid ultrasound, which measures Intima-Media Thickness (IMT) and carotid artery stenosis, among other indicators. Additionally, we calculated the AIP. This approach allowed us to thoroughly analyze the correlation between these key indicators and the severity of coronary heart disease lesions. Results: The study included 59 patients, 38 with single-vessel disease and 21 with multi-vessel disease. In the multivessel disease group, we observed significantly elevated levels of AIP, IMT, and carotid stenosis compared to the single-vessel disease group. Specifically, AIP, IMT, and carotid stenosis levels were higher in the multi-vessel group. Furthermore, our analysis revealed a positive correlation between AIP and IMT (r = 0.038, P = .003), while no significant correlation was found between AIP and carotid stenosis. Additionally, there was a moderate correlation between IMT and carotid stenosis. Conclusion: The combined assessment of AIP and carotid ultrasonography emerges as a promising approach for evaluating the severity of CHD. Notably, the multi-vessel disease group exhibited higher AIP levels compared to the single-vessel disease group, along with increased IMT and carotid artery stenosis. Our findings highlight a positive correlation between AIP and IMT, as well as between IMT and the degree of carotid stenosis. These associations underscore the potential of AIP, in conjunction with carotid ultrasonography parameters, as valuable indicators for gauging CHD severity. The clinical implications of these findings warrant further exploration, particularly in their potential integration into existing diagnostic or prognostic models for CHD. This integrated approach may offer enhanced precision in distinguishing between single-vessel and multi-vessel disease, contributing to more informed clinical decision-making.

Deep learning-based speech analysis for Alzheimer's disease detection: a literature review.

BACKGROUND: Alzheimer's disease has become one of the most common neurodegenerative diseases worldwide, which seriously affects the health of the elderly. Early detection and intervention are the most effective prevention methods currently. Compared with traditional detection methods such as traditional scale tests, electroencephalograms, and magnetic resonance imaging, speech analysis is more convenient for automatic large-scale Alzheimer's disease detection and has attracted extensive attention from researchers. In particular, deep learning-based speech analysis and language processing techniques for Alzheimer's disease detection have been studied and achieved impressive results. METHODS: To integrate the latest research progresses, hundreds of relevant papers from ACM, DBLP, IEEE, PubMed, Scopus, Web of Science electronic databases, and other sources were retrieved. We used these keywords for paper search: (Alzheimer OR dementia OR cognitive impairment) AND (speech OR voice OR audio) AND (deep learning OR neural network). CONCLUSIONS: Fifty-two papers were finally retained after screening. We reviewed and presented the speech databases, deep learning methods, and model performances of these studies. In the end, we pointed out the mainstreams and limitations in the current studies and provided a direction for future research.

A Study on Voice Measures in Patients With Alzheimer's Disease.

OBJECTIVE: As Alzheimer's disease (AD) might provoke certain nerve disorders, patients with AD can acquire sensorimotor adaptation problems, and thus the acoustic characteristics of the speech they produce may differ from those of healthy subjects. This study aimed to (1) extract acoustic characteristics (relating to articulatory gestures) potentially useful for detecting AD and (2) examine whether these characteristics could help identify AD patients. METHODS: A total of 50 individuals participated in the study, including the AD group (17 cases), the Neurologically Healthy (NH) group (13 cases), the Mild Cognitive Impairment (MCI) group (11 cases), and the Vascular Cognitive Impairment (VCI) group (9 cases). Voice samples involving three vowels (/i/, /a/, and /u/) and six consonants (/p/, /pÊ°/, /t/, /tÊ°/, /k/, and /kÊ°/) were collected using a digital recorder (TASCAM DR40X). Microphone-to-mouth distance was maintained at 30 cm. Acoustic measures included F0, jitter, shimmer, HNR, F1, F2, F3, and VOT. RESULTS: One-way ANOVA tests were carried out to compare the acoustic measures among the four groups. F3 of vowel /u/, F2 bandwidth of vowel /a/, VOT of consonant /t/, and male participants' F0 of three vowels (/a/, /i/, and /u/) were found significantly different, while no significant differences were found in the other measures. CONCLUSION: Some acoustic characteristics can indeed help detect AD patients.

Fully automated discrimination of Alzheimer's disease using resting-state electroencephalography signals.

BACKGROUND: The Alzheimer's disease (AD) population increases worldwide, placing a heavy burden on the economy and society. Presently, there is no cure for AD. Developing a convenient method of screening for AD and mild cognitive impairment (MCI) could enable early intervention, thus slowing down the progress of the disease and enabling better overall disease management. METHODS: In the current study, resting-state electroencephalography (EEG) data were acquired from 113 normal cognition (NC) subjects, 116 amnestic MCI patients, and 72 probable AD patients. After preprocessing by an automatic algorithm, features including spectral power, complexity, and functional connectivity were extracted, and machine-learning classifiers were built to differentiate among the 3 groups. The classification performance was evaluated from multiple perspectives, including accuracy, specificity, sensitivity, area under the curve (AUC) with 95% confidence intervals, and compared to the empirical chance level by permutation tests. RESULTS: The analysis of variance results (P<0.05 with false discovery rate correction) confirmed the tendency to slow brain activity, reduced complexity, and connectivity with AD progress. By combining the features, the ability of the machine-learning classifiers, especially the ensemble trees, to differentiate among the 3 groups, was significantly better than that of the empirical chance level of the permutation test. The AUC of the classifier with the best performance was 80.08% for AD vs. NC, 70.82% for AD vs. MCI, and 63.95% for MCI vs. NC. CONCLUSIONS: The current study presented a fully automatic procedure that could significantly distinguish NC, MCI, and AD subjects via resting-state EEG signals. The study was based on a large data set with evidence-based medical diagnosis and provided further evidence that resting-state EEG data could assist in the discrimination of AD patients.

Selecting and Analyzing Speech Features for the Screening of Mild Cognitive Impairment.

The total number of patients with Alzheimer's Disease (AD) has exceeded 10 million in China, while the consultation rate is only 14%. Large-scale early screening of cognitive impairment is necessary, however, the methods of traditional screening are expensive and time-consuming. This study explores a speech-based method for the early screening of cognitive impairment by selecting and analyzing speech features to reduce cost and increase efficiency. Specifically, speech-based early screening models are built based on a feature selection method and a self-built dataset including AD patients, Mild Cognitive Impairment (MCI) patients, and healthy controls. This method achieves 10% relative improvement in F1-score to discriminate MCI patients from healthy controls on our dataset. The prediction F1-score reached 70.73% when discriminating MCI patients from healthy controls based on the feature importance list calculated by the auxiliary model that is built to discriminate AD from Control group. Besides, to further assist the medical screening of MCI, we analyze the correlation between brain atrophy features and speech features including acoustic, lexical and duration features. On the basis of key speech feature selection and correlation analysis, the reference interval of speech features is constructed based on the speech data from Control group to provide a reference for evaluating cognitive impairment.Clinical Relevance - We build a speech-based dataset including AD, MCI and Control groups, and provide a feature selection method to improve the effectiveness of the screening of MCI. Apart from this, the correlation between speech features and brain atrophy features is analyzed. Finally, the reference interval of key speech features is established.

Bidirectional Attention for Text-Dependent Speaker Verification.

Automatic speaker verification provides a flexible and effective way for biometric authentication. Previous deep learning-based methods have demonstrated promising results, whereas a few problems still require better solutions. In prior works examining speaker discriminative neural networks, the speaker representation of the target speaker is regarded as a fixed one when comparing with utterances from different speakers, and the joint information between enrollment and evaluation utterances is ignored. In this paper, we propose to combine CNN-based feature learning with a bidirectional attention mechanism to achieve better performance with only one enrollment utterance. The evaluation-enrollment joint information is exploited to provide interactive features through bidirectional attention. In addition, we introduce one individual cost function to identify the phonetic contents, which contributes to calculating the attention score more specifically. These interactive features are complementary to the constant ones, which are extracted from individual speakers separately and do not vary with the evaluation utterances. The proposed method archived a competitive equal error rate of 6.26% on the internal "DAN DAN NI HAO" benchmark dataset with 1250 utterances and outperformed various baseline methods, including the traditional i-vector/PLDA, d-vector, self-attention, and sequence-to-sequence attention models.

Detecting Alzheimer's Disease from Continuous Speech Using Language Models.

BACKGROUND: Recently, many studies have been carried out to detect Alzheimer's disease (AD) from continuous speech by linguistic analysis and modeling. However, few of them utilize language models (LMs) to extract linguistic features and to investigate the lexical-level differences between AD and healthy speech. OBJECTIVE: Our goals include obtaining state-of-art performance of automatic AD detection, emphasizing N-gram LMs as powerful tools for distinguishing AD patients' narratives from those of healthy controls, and discovering the differences of lexical usages between AD patients and healthy people. METHOD: We utilize a subset of the DementiaBank corpus, including 242 control samples from 99 control participants and 256 AD samples from 169 "PossibleAD" or "ProbableAD" participants. Baseline models are built through area under curve-based feature selection and using five machine learning algorithms for comparison. Perplexity features are extracted using LMs to build enhanced detection models. Finally, the differences of lexical usages between AD patients and healthy people are investigated by a proportion test based on unigram probabilities. RESULTS: Our baseline model obtains a detection accuracy of 80.7%. This accuracy increases to 85.4% after integrating the perplexity features derived from LMs. Further investigations show that AD patients tend to use more general, less informative, and less accurate words to describe characters and actions than healthy controls. CONCLUSION: The perplexity features extracted by LMs can benefit the automatic AD detection from continuous speech. There exist lexical-level differences between AD and healthy speech that can be captured by statistical N-gram LMs.

Automatic detection and classification of marmoset vocalizations using deep and recurrent neural networks.

This paper investigates the methods to detect and classify marmoset vocalizations automatically using a large data set of marmoset vocalizations and deep learning techniques. For vocalization detection, neural networks-based methods, including deep neural network (DNN) and recurrent neural network with long short-term memory units, are designed and compared against a conventional rule-based detection method. For vocalization classification, three different classification algorithms are compared, including a support vector machine (SVM), DNN, and long short-term memory recurrent neural networks (LSTM-RNNs). A 1500-min audio data set containing recordings from four pairs of marmoset twins and manual annotations is employed for experiments. Two test sets are built according to whether the test samples are produced by the marmosets in the training set (test set I) or not (test set II). Experimental results show that the LSTM-RNN-based detection method outperformed others and achieved 0.92% and 1.67% frame error rate on these two test sets. Furthermore, the deep learning models obtained higher classification accuracy than the SVM model, which was 95.60% and 91.67% on the two test sets, respectively.

Representational similarity analysis reveals task-dependent semantic influence of the visual word form area.

Access to semantic information of visual word forms is a key component of reading comprehension. In this study, we examined the involvement of the visual word form area (VWFA) in this process by investigating whether and how the activity patterns of the VWFA are influenced by semantic information during semantic tasks. We asked participants to perform two semantic tasks - taxonomic or thematic categorization - on visual words while obtaining the blood-oxygen-level-dependent (BOLD) fMRI responses to each word. Representational similarity analysis with four types of semantic relations (taxonomic, thematic, subjective semantic rating and word2vec) revealed that neural activity patterns of the VWFA were associated with taxonomic information only in the taxonomic task, with thematic information only in the thematic task and with the composite semantic information measured by word2vec in both semantic tasks. Furthermore, the semantic information in the VWFA cannot be explained by confounding factors including orthographic, low-level visual and phonological information. These findings provide positive evidence for the presence of both orthographic and task-relevant semantic information in the VWFA and have significant implications for the neurobiological basis of reading.

Organizational Principles of Abstract Words in the Human Brain.

words constitute nearly half of the human lexicon and are critically associated with human abstract thoughts, yet little is known about how they are represented in the brain. We tested the neural basis of 2 classical cognitive notions of abstract meaning representation: by linguistic contexts and by semantic features. We collected fMRI BOLD responses for 360 abstract words and built theoretical representational models from state-of-the-art corpus-based natural language processing models and behavioral ratings of semantic features. Representational similarity analyses revealed that both linguistic contextual and semantic feature similarity affected the representation of abstract concepts, but in distinct neural levels. The corpus-based similarity was coded in the high-level linguistic processing system, whereas semantic feature information was reflected in distributed brain regions and in the principal component space derived from whole-brain activation patterns. These findings highlight the multidimensional organization and the neural dissociation between linguistic contextual and featural aspects of abstract concepts.

Human respiratory syncytial virus nonstructural protein NS2 antagonizes the activation of beta interferon transcription by interacting with RIG-I.

A wide variety of RNA viruses have been shown to produce proteins that inhibit interferon (IFN) production and signaling. For human respiratory syncytial virus (RSV), the nonstructural NS1 and NS2 proteins have been shown to block IFN signaling by causing the proteasomal degradation of STAT2. In addition, recombinant RSVs lacking either NS1 or NS2 induce more IFN production than wild-type (wt) RSV in infected cells. However, the mechanisms by which the NS proteins perform this function are unknown. In this study, we focused on defining the mechanism by which NS2 inhibits the induction of IFN transcription. We find that NS2 is required for the early inhibition of IFN transcription since the infection of cells with NS2-deletion RSV resulted in a higher level of IRF3 activation at early time points postinfection compared with that of wt or NS1-deletion RSV infection. In addition, NS2 expression inhibits IFN transcription induced by both the RIG-I and TLR3 pathways. Furthermore, we show that NS2 inhibits RIG-I-mediated IFN promoter activation by binding to the N-terminal CARD of RIG-I and inhibiting its interaction with the downstream component MAVS (IPS-1, VISA, Cardif). Thus, the RSV NS2 protein is a multifunctional IFN antagonist that targets specific components of both the IFN induction and IFN signaling pathways.

Purification and characterization of recombinant human respiratory syncytial virus nonstructural protein NS1.

We report here the first biochemical and structural characterization of the respiratory syncytial virus (RSV) NS1 protein. We have used a pET-ubiquitin expression system to produce respiratory syncytial virus (RSV) NS1 protein in E. coli that contains a hexahistidine-tag on either the amino- or carboxyl-terminus (His(6)-NS1 and NS1-His(6), respectively). We have been able to isolate milligram quantities of highly purified His(6)-NS1 and NS1-His(6) by nickel affinity chromatography. Generation of recombinant RSV indicated that addition of the hexahistidine tag to the C-terminus of NS1 slightly decreased viral replication competence whereas addition of the tag to the N-terminus had no observable effect. Therefore, we performed a comprehensive biochemical and biophysical characterization on His(6)-NS1. His(6)-NS1 is monodisperse in solution as determined by dynamic light scattering analysis. Both gel filtration and analytical ultracentrifugation showed that His(6)-NS1 is predominantly a monomer. In agreement with theoretical predictions, circular dichroism spectroscopy showed that His(6)-NS1 contains 21% alpha-helices, 34% beta-sheets, and 45% undefined structure. Immunization with purified His(6)-NS1 generated an antiserum that specifically recognizes NS1 by immunoprecipitation from HEp-2 cells infected by RSV, indicating that His(6)-NS1 resembles native NS1. The availability of purified RSV NS1 will permit biochemical and structural investigations providing insight into the function of NS1 in viral replication and interferon antagonism.