Essentials: The Science of Learning & Speaking Languages | Dr. Eddie Chang

Introduction to Speech and Language Neurobiology

• The discussion focuses on the neurobiology of speech and language, with the goal of understanding how the brain processes these complex functions, and the main difference between speech and language is that speech corresponds to the communication signal, while language refers to the meaning and context of the words being spoken 10s.
• Language is a broader concept that encompasses pragmatics, semantics, and syntax, which are all critical parts of understanding the meaning and context of words and sentences, and speech is just one form of language, with other forms including sign language and reading 2m6s.

Anatomy and Physiology of Speech Production

• The production of speech involves the coordination of multiple brain areas and physical structures, including the vocal tract, lips, tongue, ferinx, and larynx, which work together to shape the breath and produce sound, with the larynx playing a key role in bringing the vocal folds together to vibrate and produce sound 4m30s.
• The larynx is responsible for vibrating the vocal folds, also known as vocal cords, at high frequencies, typically between 100 to 200 hertz, and the size and shape of the larynx can affect the quality of a person's voice, with men and women generally having different voice qualities due to differences in larynx size and shape 6m40s.
• The process of speaking is a complex motor function that involves the coordination of multiple physical structures and brain areas, and is considered one of the most complex motor functions that humans perform, requiring precise control over the vocal tract, lips, tongue, and other structures to produce intelligible speech 5m30s.

Voice Production and Gender Differences

• The human voice is generated by the larynx, with men having a larger voice box and a lower vibrating frequency of around 100 hertz, while women have a higher frequency of around 200 hertz, and this energy is then shaped by the vocal tract to create consonants and vowels 10s.
• Vocalizations, such as crying or laughter, are different from spoken language and involve the exhalation of air and phonation at the level of the larynx, but are produced by unique neural structures that are distinct from the language areas of the brain 2m6s.
• People who have injuries in the speech and language areas of the brain can still produce vocalizations, such as moaning, and this suggests that vocalization is a different form of communication that is specialized in non-human primates and other areas of the brain 2m6s.

Brain-Machine Interfaces for Communication

• Brain machine interfaces are being used to help people who are paralyzed or "locked in" to communicate, by translating the neural activity of areas of the brain that produce speech into hardware that can artificial produce speech 4m30s.
• Conditions such as brain stem stroke, neurodegenerative diseases, and ALS can cause severe forms of paralysis, including locked-in syndrome, where a person has completely intact cognition and awareness but is unable to express themselves through speech or voluntary movement 6m0s.
• Researchers have been studying the patterning of electrical activity for consonants and vowels, and are working to develop brain machine interfaces that can help people with paralysis to communicate by decoding the neural activity associated with speech 8m0s.

Clinical Trial and Participant Background

• The goal of a clinical trial, known as the Bravo trial, is to intercept brain signals from the cerebral cortex as someone tries to speak and translate them into words using a computer, with the first participant being a man who was paralyzed for 15 years after a car accident and a subsequent stroke in the brain stem 10s.
• The participant, who had been unable to speak or move his arms and legs for 15 years, could only blink his eyes and had limited mouth movements, but no intelligible speech, and had developed a way to communicate through a device that allowed him to type out letters on a keyboard screen by moving his neck 2m6s.
• The participant underwent a brain surgery where electrodes were implanted onto areas that control the vocal tract, larynx, lips, tongue, and jaw movements, and an electrode array was connected to a port that was screwed to his skull, allowing him to live with the implant for the last three years 4m30s.

Technology and Algorithm Development

• The electrode array is connected to a wire that translates brain waves into digital signals, which are then put through a machine learning algorithm that can pick up subtle patterns and translate them into words, with the algorithm taking weeks to train and initially limited to a vocabulary of 50 words 6m40s.
• The participant's reaction to being able to communicate through the device was emotional, with his body and head shaking and giggling, but this also affected the decoding of the next words, and while this issue has not been fixed, the participant is simply told to stop giggling 10m20s.
• The vocabulary used in the trial is expanding, with the initial 50 words allowing for the creation of possible sentences, and it is expected that the vocabulary will become much larger over time 12m30s.

Error Correction and Future Improvements

• The creation of a computational model of word combinations can be used to develop an autocorrect feature, similar to those used in texting, which can correct errors in decoding brain activity patterns, allowing paralyzed individuals to create words and sentences 10s.
• Brain machine interface technology has been in development for decades, with early work focusing on restoring arm movement or controlling computer cursors, and has now started to become commercialized with potential medical applications 2m6s.

Ethical and Historical Considerations of Augmentation

• The concept of augmentation, or enhancing human abilities beyond normal levels, is a serious area of consideration, with potential applications including superior memory, communication speeds, and athletic abilities, but also raises important ethical questions 2m6s.
• The pursuit of augmentation is not new, as humans have been using various methods to enhance their abilities throughout history, such as coffee, nicotine, and medications, but the use of neurotechnologies raises unique questions about invasive procedures and non-medical applications 2m6s.
• The development of brain machine interface technology is likely to continue, with potential applications emerging in incremental and subtle ways, such as enhanced cognitive abilities, rather than dramatic or obvious changes 2m6s.

Evolution of Communication and Technological Limitations

• The current systems for speaking and communicating have evolved over thousands and millions of years, supported by neural structures with a bandwidth of millions of neurons, and no existing technology can match this capability 10s.
• The topic of augmentation, both physical and cognitive, is not new, as humans have already surpassed this, and the focus is now on enhanced cognition, where technology will be the rate-limiting step, raising questions about its impact on society and access to this technology 42s.

Nonverbal Communication and Avatars

• Nonverbal expressions, such as facial expressions, are a crucial part of communication, as they provide feedback and allow for adjustments in speech, and seeing the mouth move and jaw move can improve intelligibility and make communication more natural 2m6s.
• Researchers are working on merging brain-machine interfaces with extraction of speech signals from people who are locked in, like Poncho, to improve their ability to communicate, and one approach is building holistic avatars that can decode expressions and movements associated with speech 4m10s.
• The goal is to create a system where people can communicate through a fully computer-animated face, or avatar, that can convey speech movements and facial expressions, making communication more complete and natural, and this technology is expected to become a reality soon 6m30s.
• This technology has important implications for people with disabilities, as it can help them participate in digital and virtual social interactions, and researchers are working to make this technology accessible to those who need it, with the potential for avatars to speak on behalf of individuals, revolutionizing the way people communicate 8m40s.

Speech Neuroprosthetics and Embodied Communication

• A speech neuroprosthetic is a device that can help people learn to speak by providing feedback, allowing users to embody and feel like they are directly controlling an avatar or animation, which can be an effective way to learn quickly 10s.
• Stuttering is a speech condition where individuals have difficulty speaking fluently, despite having the necessary language skills, and it affects articulation, specifically the coordinated movements in the vocal tract required to produce fluent speech 42s.

Understanding and Treating Stuttering

• Stuttering is a condition that people can be predisposed to, but it is not present all the time, and anxiety can trigger or worsen stuttering, although it is not the primary cause, which is still not fully understood but is related to brain functions 1m6s.
• The cause of stuttering is linked to a breakdown in the complex and precise machinery of speech production, which is typically unconscious and naturally developed, and it can be thought of as a disruption in the coordination of the areas controlling the vocal tract, similar to a symphony 2m6s.
• Addressing stuttering early in life, when neuroplasticity is robust, can be beneficial, and typical treatment involves speech therapy, which may include working through anxiety and developing tricks to create conditions that allow for fluent speech, such as focusing on initiation problems and using auditory feedback 4m10s.
• Auditory feedback, which is the sound of one's own voice, plays a crucial role in speech production, and altering it can impact stuttering, making it either better or worse, and speech therapists can help individuals work through these issues 5m40s.

Neurological Coordination and Speech Fluency

• The brain is not only focused on sending commands, but also interacts with the part that hears sounds, and this connection may break down when stuttering occurs, indicating a precise coordination in the brain is necessary for fluent speech 10s.
• There are individuals who stutter, but not all the time, suggesting that in moments of fluent speech, something specific happens in the brain to enable this precise coordination 42s.

Personal Reflection and Conclusion

• A personal anecdote is shared about a 38-year friendship, and the person is thrilled to be sitting with Eddie, appreciating his work on the cutting edge of medicine and science, including his role as a chair of a department working alongside others doing incredible work 2m6s.