Artificial Neural Networks Combined with the Principal Component Analysis for Non-Fluent Speech Recognition.
Ontology highlight
ABSTRACT:
SUBMITTER: Swietlicka I
PROVIDER: S-EPMC8749906 | biostudies-literature |
REPOSITORIES: biostudies-literature
Similar Datasets
Project description:IntroductionIn recent years, machines powered by deep learning have achieved near-human levels of performance in speech recognition. The fields of artificial intelligence and cognitive neuroscience have finally reached a similar level of performance, despite their huge differences in implementation, and so deep learning models can-in principle-serve as candidates for mechanistic models of the human auditory system.MethodsUtilizing high-performance automatic speech recognition systems, and advanced non-invasive human neuroimaging technology such as magnetoencephalography and multivariate pattern-information analysis, the current study aimed to relate machine-learned representations of speech to recorded human brain representations of the same speech.ResultsIn one direction, we found a quasi-hierarchical functional organization in human auditory cortex qualitatively matched with the hidden layers of deep artificial neural networks trained as part of an automatic speech recognizer. In the reverse direction, we modified the hidden layer organization of the artificial neural network based on neural activation patterns in human brains. The result was a substantial improvement in word recognition accuracy and learned speech representations.DiscussionWe have demonstrated that artificial and brain neural networks can be mutually informative in the domain of speech recognition.
| S-EPMC9811675 | biostudies-literature
Project description:The nature and frequency of speech production errors in neurodegenerative disease have not previously been precisely quantified. In the present study, 16 patients with a progressive form of non-fluent aphasia (PNFA) were asked to tell a story from a wordless children's picture book. Errors in production were classified as either phonemic, involving language-based deformations that nevertheless result in possible sequences of English speech segments; or phonetic, involving a motor planning deficit and resulting in non-English speech segments. The distribution of cortical atrophy as revealed by structural MRI scans was examined quantitatively in a subset of PNFA patients (N=7). The few errors made by healthy seniors were only phonemic in type. PNFA patients made more than four times as many errors as controls. This included both phonemic and phonetic errors, with a preponderance of errors (82%) classified as phonemic. The majority of phonemic errors were substitutions that shared most distinctive features with the target phoneme. The systematic nature of these substitutions is not consistent with a motor planning deficit. Cortical atrophy was found in prefrontal regions bilaterally and peri-Sylvian regions of the left hemisphere. We conclude that the speech errors produced by PNFA patients are mainly errors at the phonemic level of language processing and are not caused by a motor planning impairment.
| S-EPMC2839014 | biostudies-literature
Project description:Artificial neural networks (ANN) have become the mainstream acoustic modeling technique for large vocabulary automatic speech recognition (ASR). A conventional ANN features a multi-layer architecture that requires massive amounts of computation. The brain-inspired spiking neural networks (SNN) closely mimic the biological neural networks and can operate on low-power neuromorphic hardware with spike-based computation. Motivated by their unprecedented energy-efficiency and rapid information processing capability, we explore the use of SNNs for speech recognition. In this work, we use SNNs for acoustic modeling and evaluate their performance on several large vocabulary recognition scenarios. The experimental results demonstrate competitive ASR accuracies to their ANN counterparts, while require only 10 algorithmic time steps and as low as 0.68 times total synaptic operations to classify each audio frame. Integrating the algorithmic power of deep SNNs with energy-efficient neuromorphic hardware, therefore, offer an attractive solution for ASR applications running locally on mobile and embedded devices.
| S-EPMC7090229 | biostudies-literature
Project description:Deficits in fluent speech production following left hemisphere stroke are a central concern because of their impact on patients' lives and the insight they provide about the neural organization of language processing. Fluent speech production requires the rapid coordination of phonological, semantic, and syntactic processing, so this study examined how deficits in connected speech relate to these language sub-systems. Behavioural data (N = 69 participants with aphasia following left hemisphere stroke) consisted of a diverse and comprehensive set of narrative speech production measures and measures of overall severity, semantic deficits, and phonological deficits. These measures were entered into a principal component analysis with bifactor rotation-a latent structure model where each item loads on a general factor that reflects what is common among the items, and orthogonal factors that explain variance not accounted for by the general factor. Lesion data were available for 58 of the participants, and each factor score was analysed with multivariate lesion-symptom mapping. Effects of connectivity disruption were evaluated using robust regression with tract disconnection or graph theoretic measures of connectivity as predictors. The principal component analysis produced a four-factor solution that accounted for 70.6% of the variance in the data, with a general factor corresponding to the overall severity and length and complexity of speech output (complexity factor), a lexical syntax factor, and independent factors for Semantics and Phonology. Deficits in the complexity of speech output were associated with a large temporo-parietal region, similar to overall aphasia severity. The lexical syntax factor was associated with damage in a relatively small set of fronto-parietal regions which may reflect the recruitment of control systems to support retrieval and correct usage of lexical items that primarily serve a syntactic rather than semantic function. Tract-based measures of connectivity disruption were not statistically associated with the deficit scores after controlling for overall lesion volume. Language network efficiency and average clustering coefficient within the language network were significantly associated with deficit scores after controlling for overall lesion volume. These results highlight overall severity as the critical contributor to fluent speech in post-stroke aphasia, with a dissociable factor corresponding to lexical syntax.
| S-EPMC9798301 | biostudies-literature
Project description:The potential of Fourier Transform infrared microspectroscopy (FTIR microspectroscopy) and multivariate analyses were applied for the classification of the frequency ranges responsible for the distribution changes of the main components of articular cartilage (AC) that occur during dietary β-hydroxy-β-methyl butyrate (HMB) supplementation. The FTIR imaging analysis of histological AC sections originating from 35-day old male piglets showed the change in the collagen and proteoglycan contents of the HMB-supplemented group compared to the control. The relative amount of collagen content in the superficial zone increased by more than 23% and in the middle zone by about 17%, while no changes in the deep zone were observed compared to the control group. Considering proteoglycans content, a significant increase was registered in the middle and deep zones, respectively; 62% and 52% compared to the control. AFM nanoindentation measurements collected from animals administered with HMB displayed an increase in AC tissue stiffness by detecting a higher value of Young's modulus in all investigated AC zones. We demonstrated that principal component analysis and artificial neural networks could be trained with spectral information to distinguish AC histological sections and the group under study accurately. This work may support the use and effectiveness of FTIR imaging combined with multivariate analyses as a quantitative alternative to traditional collagenous tissue-related histology.
| S-EPMC8430473 | biostudies-literature
Project description:Patients with non-fluent aphasias display impairments of expressive and receptive grammar. This has been attributed to deficits in processing configurational and hierarchical sequencing relationships. This hypothesis had not been formally tested. It was also controversial whether impairments are specific to language, or reflect domain general deficits in processing structured auditory sequences. Here we used an artificial grammar learning paradigm to compare the abilities of controls to participants with agrammatic aphasia of two different aetiologies: stroke and frontotemporal dementia. Ten patients with non-fluent variant primary progressive aphasia (nfvPPA), 12 with non-fluent aphasia due to stroke, and 11 controls implicitly learned a novel mixed-complexity artificial grammar designed to assess processing of increasingly complex sequencing relationships. We compared response profiles for otherwise identical sequences of speech tokens (nonsense words) and tone sweeps. In all three groups the ability to detect grammatical violations varied with sequence complexity, with performance improving over time and being better for adjacent than non-adjacent relationships. Patients performed less well than controls overall, and this was related more strongly to aphasia severity than to aetiology. All groups improved with practice and performed well at a control task of detecting oddball nonwords. Crucially, group differences did not interact with sequence complexity, demonstrating that aphasic patients were not disproportionately impaired on complex structures. Hierarchical cluster analysis revealed that response patterns were very similar across all three groups, but very different between the nonsense word and tone tasks, despite identical artificial grammar structures. Overall, we demonstrate that agrammatic aphasics of two different aetiologies are not disproportionately impaired on complex sequencing relationships, and that the learning of phonological and non-linguistic sequences occurs independently. The similarity of profiles of discriminatory abilities and rule learning across groups suggests that insights from previous studies of implicit sequence learning in vascular aphasia are likely to prove applicable in nfvPPA.
| S-EPMC5637161 | biostudies-literature
Project description:ObjectiveHigh-frequency repetitive transcranial magnetic stimulation (HF-rTMS) to the lesional hemisphere requires prudence in selecting the appropriate stimulation spot. Functional near-IR spectroscopy (fNIRS) can be used in both selecting the stimulation spot and assessing the changes of the brain network. This study aimed to evaluate the effect of HF-rTMS on the most activated spot identified with fNIRS and assess the changes of brain functional network in the patients with poststroke aphasia.MethodsA total of five patients received HF-rTMS to the most activated area on the lesional hemisphere, followed by 30 min of speech therapy for 10 days. The Korean version of the Western aphasia battery (K-WAB) and fNIRS evaluation were done 1 day before the treatment, 1 day and 1 month after the last treatment session. Changes of K-WAB and paired cortical interaction and brain network analysis using graph theory were assessed.ResultsAphasia quotient in K-WAB significantly increased after the treatment (P = 0.043). The correlation analysis of cortical interactions showed increased connectivity between language production and processing areas. Clustering coefficients of the left hemisphere were increased over a sparsity range between 0.45 and 0.58 (0.015 < p < 0.031), whereas the clustering coefficients of the right hemisphere, decreased over a sparsity range 0.15-0.87 (0.063 < p < 0.095). The global efficiency became lower over a network sparsity range between 0.47 and 0.75 (0.015 < p < 0.063).ConclusionImprovement of language function and changes of corticocortical interaction between language-related cortical areas were observed after HF-rTMS on the most activated area identified by fNIRS with combined speech therapy in the patients with poststroke aphasia.
| S-EPMC8866644 | biostudies-literature
Project description:The non-fluent/agrammatic variant of primary progressive aphasia (nfvPPA) presents with a gradual decline in grammar and motor speech resulting from selective degeneration of speech-language regions in the brain. There has been considerable progress in identifying treatment approaches to remediate language deficits in other primary progressive aphasia variants; however, interventions for the core deficits in nfvPPA have yet to be systematically investigated. Further, the neural mechanisms that support behavioural restitution in the context of neurodegeneration are not well understood. We examined the immediate and long-term benefits of video implemented script training for aphasia (VISTA) in 10 individuals with nfvPPA. The treatment approach involved repeated rehearsal of individualized scripts via structured treatment with a clinician as well as intensive home practice with an audiovisual model using 'speech entrainment'. We evaluated accuracy of script production as well as overall intelligibility and grammaticality for trained and untrained scripts. These measures and standardized test scores were collected at post-treatment and 3-, 6-, and 12-month follow-up visits. Treatment resulted in significant improvement in production of correct, intelligible scripted words for trained topics, a reduction in grammatical errors for trained topics, and an overall increase in intelligibility for trained as well as untrained topics at post-treatment. Follow-up testing revealed maintenance of gains for trained scripts up to 1 year post-treatment on the primary outcome measure. Performance on untrained scripts and standardized tests remained relatively stable during the follow-up period, indicating that treatment helped to stabilize speech and language despite disease progression. To identify neural predictors of responsiveness to intervention, we examined treatment effect sizes relative to grey matter volumes in regions of interest derived from a previously identified speech production network. Regions of significant atrophy within this network included bilateral inferior frontal cortices and supplementary motor area as well as left striatum. Volumes in a left middle/inferior temporal region of interest were significantly correlated with the magnitude of treatment effects. This region, which was relatively spared anatomically in nfvPPA patients, has been implicated in syntactic production as well as visuo-motor facilitation of speech. This is the first group study to document the benefits of behavioural intervention that targets both linguistic and motoric deficits in nfvPPA. Findings indicate that behavioural intervention may result in lasting and generalized improvement of communicative function in individuals with neurodegenerative disease and that the integrity of spared regions within the speech-language network may be an important predictor of treatment response.
| S-EPMC5972572 | biostudies-literature
Project description:Seasonally frozen soils are exposed to freeze‒thaw cycles every year, leading to mechanical property deterioration. To reasonably describe the deterioration of soil under different conditions, machine learning (ML) technology is used to establish a prediction model for soil static strength. Six key influencing factors (moisture content, compaction degree, confining pressure, freezing temperature, number of freeze‒thaw cycles and thawing duration) are included in the modelling database. The accuracy of three typical ML algorithms (support vector machine (SVM), random forest (RF) and artificial neural network (ANN)) is compared. The results show that the ANN outperforms the SVM and RF. Principal component analysis (PCA) is combined with the ANN, and the PCA-ANN algorithm is proposed, which further improves the prediction accuracy. The deterioration of soil static strength is systematically researched using the PCA-ANN algorithm. The results show that the soil static strength decreased considerably after the first several freeze‒thaw cycles before the strength plateau occurred, and the strength reduction increased significantly with increasing moisture content and compaction degree. The PCA-ANN model can generate a reasonable prediction for the static strength or other soil properties of seasonally frozen soil, which will provide a scientific reference for practical engineering.
| S-EPMC10522631 | biostudies-literature
Project description:Neurons in higher cortical areas, such as the prefrontal cortex, are often tuned to a variety of sensory and motor variables, and are therefore said to display mixed selectivity. This complexity of single neuron responses can obscure what information these areas represent and how it is represented. Here we demonstrate the advantages of a new dimensionality reduction technique, demixed principal component analysis (dPCA), that decomposes population activity into a few components. In addition to systematically capturing the majority of the variance of the data, dPCA also exposes the dependence of the neural representation on task parameters such as stimuli, decisions, or rewards. To illustrate our method we reanalyze population data from four datasets comprising different species, different cortical areas and different experimental tasks. In each case, dPCA provides a concise way of visualizing the data that summarizes the task-dependent features of the population response in a single figure.
| S-EPMC4887222 | biostudies-literature