Magnetic Fields of Speech: Russian Scientists Develop Brain-Based Method for Diagnosing Speech Disorders
Russian researchers have developed a method that tracks how the brain reacts to unexpected words in natural speech using magnetoencephalography. The technology could help identify dyslexia in children at earlier stages.
Russian neurophysiologists from HSE University have developed an approach that makes it possible to track the brain’s response to unexpected words in ordinary, continuous speech. The method uses magnetoencephalography, or MEG, which records the magnetic fields generated by neural activity.
Traditional evoked-potential methods work differently. Participants repeatedly hear the same word while researchers record electrical brain signals and average the results. That approach is of limited use for natural speech. In real-life communication, the brain does not process isolated sounds one by one. Instead, it continuously constructs meaning, predicts upcoming words and links speech to contextual cues. Repeating the same phrase removes that dynamic process.
A New Approach to the Problem
Russian researchers adapted the method for continuous text streams. A volunteer listens to an audio recording, for example a popular-science story about animals. Using MEG, scientists record the brain’s response to every word and then analyse the signals according to semantic categories, distinguishing words that fit the context from those that are semantically distant or unexpected. Language-based neural networks are used for the analysis.
Testing involving 27 healthy volunteers confirmed that the brain responds differently to predictable and unexpected words within a speech stream. Researchers were able to detect and measure that difference.
A Key Tool for Dyslexia Diagnosis
The new technology is particularly important for dyslexia diagnostics. Dyslexia is a disorder that affects the ability to acquire reading and writing skills. In practical terms, a child or adult with dyslexia may otherwise be intelligent, capable of reasoning and strong in mathematics, yet still struggle to recognise letters, confuse their order, read slowly or fail to understand written text.
Dyslexia affects an estimated 5% to 15% of children and adults worldwide. No objective hardware-based diagnostic method currently exists. Physicians still rely primarily on behavioural tests, questionnaires and speech therapist evaluations, processes that can be lengthy, subjective and not always precise.
The Russian-developed method offers a different approach. MEG does not require patients to provide immediate active responses because the participant simply listens to a text. The system records how the brain processes semantic inconsistencies in speech.
In Russia, the technology could eventually become a large-scale tool for early diagnosis. Earlier identification means earlier intervention, increasing the likelihood that children can catch up academically with their peers and avoid long-term reading and writing difficulties.
In the international context, the significance of the Russian approach lies in its adaptation specifically for continuous speech combined with the use of semantic categories and language-based neural networks.
What the Method Could Mean for Patients
During the diagnostic process, children no longer need to spend time working through flash cards, repeating difficult words or completing lengthy tests. Instead, they simply listen to a short text while specialists record indicators of brain activity. In this case, children experience neither stress nor fear of making mistakes.
The method makes it possible not only to identify the disorder, but also to better understand its underlying mechanism. At what stage does the disruption occur? Does the brain fail to distinguish differences between words and meaning, or does it recognise the difference but fail to connect it to context? Without an objective diagnostic tool, answering those questions is difficult. With the new approach, it becomes far more achievable.
Over time, the methodology could prove useful not only for dyslexia diagnostics, but also for other speech-related disorders, including autism spectrum disorders, post-stroke rehabilitation and traumatic brain injuries.
Export Potential
Magnetoencephalography remains an expensive technology, but MEG systems are already available in major medical centres around the world. At the same time, there are relatively few software-based analytical methods adapted specifically for continuous speech. The Russian approach represents a mathematical and linguistic algorithm that could be packaged as software and commercialised as a medical technology product.
The research was supported by a grant from the Russian Science Foundation. The next stage involves collecting larger datasets and creating benchmark indicators for different age groups and types of speech disorders. That work would then support the development of clinical protocols and broader integration into routine medical practice.
The approach can also be adapted for other languages. Russian, with its flexible word order and rich morphology, provides a strong testing ground for refining such algorithms. Researchers believe that methods effective in Russian could eventually be adapted for English, German or Chinese as well.
