The Sound of Science: Connections between Neuroscience, Physics, and Music

Summary

The event explored how physics, neuroscience, and music connect: how sound is produced and analyzed, how the brain turns acoustic signals into musical perception, and whether brain activity can be used to reconstruct music.

1. Sound, Instruments, and the Physics of Hearing — Terry Buehler

• Musical sounds begin with vibrating objects—such as strings, air columns, or bells—which create pressure waves that travel to the ear or a microphone.
• A sound’s frequency spectrum shows which frequencies it contains:
  • A pure tone, such as a whistle or tuning fork, is dominated by one frequency.
  • Most instruments produce a fundamental frequency along with harmonics or overtones. Their relative strengths help create an instrument’s timbre.
  • Some instruments, such as bells, have overtones that are not integer multiples of the fundamental frequency, so their spectra are not conventionally harmonic.
• The inner ear’s cochlea acts somewhat like a frequency analyzer. Its basilar membrane responds to different frequencies at different locations: high frequencies near the narrow, stiff end, and low frequencies farther along the membrane.
• Nearby frequencies can excite overlapping regions of the basilar membrane. Depending on their spacing, listeners may hear beating, roughness, or two distinct tones. This helps explain why some combinations of notes sound dissonant.
• The relationship between notes and perceived consonance depends partly on the overtones of the instruments involved. Notes that work well together on one type of instrument may not work the same way on another.
• Musical tuning systems and preferences can reflect the instruments and traditions of a culture. Buehler contrasted Western harmonic instruments and the 12-tone scale with Pacific Island instruments whose inharmonic overtones can support different tunings.
• In the Q&A, Buehler noted that the commonly cited human hearing range is about 20 Hz to 20,000 Hz, though high-frequency hearing often declines with age. He also explained that differences in sound-arrival time between the ears help people locate sounds.

2. How the Brain Organizes Complex Sounds — Frédéric Theunissen

• The cochlea separates sound into frequency components, but that analysis alone does not explain the unified experience of melody, timbre, rhythm, and voice.
• Theunissen described a further stage of processing in the auditory cortex, which he characterized as a kind of modulation filter bank:
  • Temporal modulations—changes in sound over time—contribute to perceptions such as rhythm.
  • Spectral modulations—patterns across frequencies—contribute to pitch and timbre.
  • Combined spectral and temporal patterns help represent features such as vibrato.
• Different cortical filters emphasize different aspects of a sound. Filtering to emphasize timbre can reduce some melodic or pitch information; emphasizing pitch can diminish the sound’s timbral quality.
• Music also involves expectations. The brain can anticipate what note or musical event may come next, and researchers have identified auditory-cortex regions whose activity is sensitive to whether musical expectations are followed or violated.
• The talk distinguished music-related expectations from speech-related expectations. Some brain regions showed sensitivity to musical patterns, while others were more responsive to expectations about speech sounds.
• Theunissen suggested that musical expectations are substantially shaped by experience and culture, much as language expectations are learned. Familiarity with a musical style can influence what a listener expects.
• He also described broad emotional tendencies in sounds: lower, more tonal sounds can be more soothing, while short, high-frequency, or dissonant sounds can signal alarm or anxiety. These patterns occur in animal communication as well as human sound perception.
• In response to questions, he noted that experience with tonal languages may be associated with somewhat better pitch discrimination but weaker rhythm discrimination. He also mentioned that plants can sense vibrations through mechanoreceptors.

3. Singing, Aphasia, and Reconstructing Sound from Brain Activity — Robert “Bob” Knight

• Knight began with the clinical observation that some people with aphasia—often caused by damage to the left hemisphere—have serious difficulty speaking but can still sing, including producing phrases they cannot say normally.
• Studying this ability may help reveal how speech and singing are represented in the brain and could eventually inform approaches to assist people with communication impairments.
• Knight described research using patients with epilepsy who had electrodes temporarily placed on or in the brain as part of clinical monitoring. These recordings let researchers compare brain activity with sounds that patients hear or produce.
• Brain-signal reconstruction approach:
  1. Present sounds, such as speech or music, while recording electrical activity from auditory brain regions.
  2. Use machine-learning methods to associate activity at different electrodes with acoustic frequencies or combinations of frequencies.
  3. Combine those frequency estimates to reconstruct an approximation of the original sound.
• The researchers have reconstructed spoken words and musical excerpts from brain recordings. Knight showed a reconstruction of Pink Floyd’s “Another Brick in the Wall”; it captured recognizable acoustic structure but sounded less clear than the original.
• In the music reconstruction work, the strongest contributions came from right-hemisphere regions, including temporal and posterior areas. This contrasts with the typical left-hemisphere dominance of language.
• In a study comparing speaking and singing the same sentence, singing produced robust activity in right-hemisphere motor and temporal regions, while left-hemisphere language areas remained active. This may help explain why singing can remain possible for some people who cannot speak.
• Knight suggested that these findings could eventually help guide treatments or communication prostheses, though the work is still at an early research stage.
• In the Q&A, he noted that singing can remain accessible to some people with aphasia caused by stroke, tumors, or degenerative conditions. He also described evidence that listeners can detect an unexpected note in a familiar tune very quickly—and that expectations can be learned even for unfamiliar music.

Closing Remarks

Steven Kahn emphasized that basic science underpins later applications and technological progress. The event’s closing message was that music offers a productive way to investigate both the physical principles of sound and the biological systems that perceive it.

Speakers and Sources Featured

• Richard Harland — Opening remarks; Dean of Biological Sciences.
• Frances Hellman — Moderator; Professor of the Graduate School in Physics and former Dean of Mathematical and Physical Sciences.
• Terry Buehler — Lecturer in Physics; presented on the physics of sound, instruments, and hearing.
• Frédéric Theunissen — Professor of Integrative Biology, Psychology, and Neuroscience; presented on auditory processing and musical expectations.
• Robert “Bob” Knight — Professor of Neuroscience and Psychology; presented on aphasia, singing, and reconstructing sound from brain activity.
• Steven Kahn — Closing remarks; Dean of Mathematical and Physical Sciences.