With a focus on human language and cognition
Open Brain AI provides online tools to enable clinicians and educators analyze written, and spoken text productions from their patients or students. The following domains are currently covered by the platform.
Large Language Models enable clinicians and teachhers by automatically analyzing spoken and written language.
We are developing automated ways to analyze spoken and written discourse productions, identify pathological characteristics, and provide scores of language and cognition.
Our research focuses on all domains of speech language and communication translatable to applications for clinicians and teachers.
Our future research and directions.
We provide discourse analysis using AI. We offer recommendations on the Cohesion and Coherence, checks the errors, and suggests whether there are pathological impairments.
We analyze written speech and provide measures and scores about all linguistic domains, including phonology, morphology, syntax, semantics, lexicon, and readability measures.
We provide automatic transcription of spoken speech and provide measures and scores about all linguistic domains, including phonology, morphology, syntax, semantics, lexicon, and readability measures..
Score the semantics of speech productions in patients, e.g., in confrontational naming tasks.
Score the phonological productions of patients.
Multilingual IPA transcription and get measures from the IPA transcription.
Computational language assessment is a fast and objective way to identify speech, language, and communication impairments as symptoms of other conditions (e.g., dementia or developmental language impairments) (Themistocleous et al. 2023).
We provide automatic online scores that can help assess the efficacy of clinical therapy or teaching (Themistocleous et al. 2021).
Language can be employed for early dementia diagnosis and deferential diagnosis into variants and prognosis (Themistocleous et al. 2022).
Providing telemonitory and telehealth with accessible tools (Themistocleous 2023).
See how your users experience your website in realtime or view trends to see any changes in performance over time.
Natural Language Processing allows the analysis of discourse automatically. It provides measures, including counts from phonology (e.g., number of syllables), morphology (e.g., number of parts of speech in text), syntax (e.g., syntactic constituents), semantics (e.g., number of entities), readability measures, and lexical measures.
It enable the transcription of texts, the automatic segmentation of speech into words and speakers, and the elicitation of acoustic measures, such as information about prosody and voice quality.
State-of-art Machine Learning and statistical models, especially Deep Neural Network architectures enable us to find patterns from language that characterize language impairments in adults and children.
Read more about our research at Charalambos Themistocleous's Personal Website
We developed a Machine Learning model to support differential diagnosis of patients with Primary Progressive Aphasia, using combined acoustic and linguistic information elicited automatically. The end-to-end automated machine learning approach enables clinicians and researchers to provide an easy, quick, and inexpensive classification of patients with PPA.
Read moreWe analyzed voice quality and speech fluency in 26 patients with Mild Cognitive Impairment (MCI) and 29 healthy controls using a speech task. We found significant differences in speech acoustic features between the two groups, suggesting voice quality, prosody, and speech analysis are an objective tool for MCI diagnosis.
Read moreUsing an automated analysis of a short picture description task, this study shows that content versus function words can distinguish patients with nonfluent PPA, semantic PPA, and logopenic PPA variants. Verbs are less important as distinguishing features of patients with different PPA variants than earlier thought. Finally, the study showed that among the most important distinguishing features of PPA variants were elaborative speech elements, such as adjectives and adverbs.
Read more