Flexible Representation of Speech

Flexible Representation of Speech in the Supratemporal Plane

The overarching goal of this exploratory research is to understand the dynamic and flexible nature of speech processing in the human supratemporal plane. The temporal lobe has long been established as a region of interest in the speech perception and processing literature because it contains the auditory cortex. More recently, research has localized the supratemporal plane as an area that exhibits response specificity to acoustic properties of complex auditory signals like speech. The supratemporal plane, comprised of Heschl's gyrus, the planum polare, and the planum temporale, is capable of the rapid spectrotemporal analysis required to map acoustic information to linguistic representation. Neural activity in this area, however, is rarely studied directly because it is difficult to access with non-invasive measures like scalp electroencephalography (EEG). Capitalizing on the unique opportunity to access these areas via routine clinical stereoelectroencephalography (sEEG) in a patient population, this study seeks to understand how cortical responses reflect the diagnosticity of two acoustic-phonetic dimensions of interest and how responses rapidly and flexibly adapt to changes in listening demands. Examining how neural response to voice onset time (VOT) and fundamental frequency (F0) modulates as a function of perceptual weight carried in signaling phoneme categories, and identifying how changes in listening context shift perceptual weight, will provide invaluable data that indicates how speech processing flexibly adapts to short-term acoustic patterns.

Study Overview

• Status: Completed
• Conditions: Epilepsy
• Intervention / Treatment: Behavioral: Dimension-Based Statistical Learning

Detailed Description

The purpose of this study is to understand the dynamic, flexible nature of speech processing as a function of perceptual weight applied to acoustic-phonetic dimensions within varying listening contexts and demands.

The specific aims of this study are as follows:

1. To establish the neural response to two acoustic-phonetic dimensions as a function of the perceptual weight they carry when signaling phoneme identity.

   Aim 1 will specifically evaluate responses to voice onset time (VOT) and fundamental frequency (F0). Data collected will provide a baseline response for participants.

2. To identify how experimental manipulation of listening context impacts perceptual weighting strategies of VOT and F0.

Aim 2 will evaluate modulation of neural response to the introduction of noise and the introduction of an "accent."

A secondary aim of this study is to use "control" electrodes, which are those placed in clinically necessary regions of the brain but outside of the region of interest for this study (supratemporal plane), to determine if additional regions of the brain are implicated in adaptive plasticity of speech processing.

Speech is the primary means by which we convey our needs, wants, and thoughts to others and the ability to process speech is crucial to our everyday functioning, as well as our ability to establish and maintain relationships. Impairments in speech processing have an undeniable negative impact on individuals and society. While habilitative and rehabilitative strategies exist that can improve auditory processing and quality of life, understanding the exact neural mechanism underlying the human brain's ability to process speech would contribute to a more well-defined means by which to target deficits. This study seeks to understand the regions of the brain involved in speech processing, how those regions analyze specific acoustic-phonetic dimensions, and how the system adapts to successfully process speech in different listening contexts.

Modern electrophysiological techniques have revolutionized research into activity in the human brain, allowing investigators to identify specific regions or patterns of activity associated with various behaviors and sensory experiences. sEEG recordings, which involve intracerebral measurements of neural activity using depth electrodes, are capable of providing unique access to regions of the brain that are otherwise inaccessible with less invasive measurements. Capitalizing on PI Abel's work with pediatric patients undergoing sEEG recording for localization of seizure foci or language mapping, this study will allow researchers to directly study activity in regions that have already been implicated in the literature as crucial to spectrotemporal analysis of complex acoustic signals like speech. These regions within the supratemporal plane (STP) include Heschl's gyrus, the planum polare, and the planum temporale, all of which are uniquely targeted via sEEG.

Existing literature in speech processing has indicated that the mapping of physical input (acoustic signal) to linguistic representation (identification of phonemes or words) is not a static process, but rather highly dependent on listening context. The auditory processing system regularly adapts to changes in signal quality, adverse listening conditions, and short-term deviations from expected and learned regularities in native language input by applying varying importance, or perceptual weight, to specific acoustic-phonetic parameters. This indicates the existence of adaptive plasticity in speech processing, yet existing neurophysiological models do not account for this flexibility in cortical response. Data from pilot EEG and sEEG studies demonstrated that high gamma activity in the STP and behavioral responses were graded by the perceptual weight given to two acoustic-phonetic dimensions, voice onset time (VOT) and fundamental frequency (F0).

The proposed study will contribute to existing knowledge by helping to establish a more detailed model of on-line cortical response and adaptation to changing acoustic signals. It is unique in its accounting for the role that perceptual weight of acoustic-phonetic dimensions play in signaling phonemes and making category-based judgments.

• Study Type: Interventional
• Enrollment (Actual): 48
• Phase: Not Applicable

Contacts and Locations

Study Locations

• United States
  • Pennsylvania
    • Pittsburgh, Pennsylvania, United States, 15224
      • UPMC Children's Hospital of Pittsburgh

Participation Criteria

Eligibility Criteria

• Ages Eligible for Study: 15 years to 25 years (Child, Adult)
• Accepts Healthy Volunteers: No

Description

Inclusion Criteria:

• Individuals 15-25 years old
• Undergoing sEEG placement in the supratemporal plane for clinically necessary localization of epileptic foci or language mapping
• Fluent English speakers
• Cognition and speech-language skills within normal limits (as determined by evaluation prior to surgery)
• Normal or correct-to-normal visual acuity
• Normal hearing acuity in each ear (as determined by audiometric assessment)
• No history of autism or ADHD

Exclusion Criteria:

• Individuals with intellectual disabilities
• Abnormal epileptiform activity in the supratemporal plane
• Lack of fluent English comprehension/production
• Severe language or auditory-specific cognitive dysfunction
• History of autism or ADHD

Study Plan

How is the study designed?

Design Details

• Primary Purpose: Basic Science
• Allocation: N/A
• Interventional Model: Single Group Assignment
• Masking: None (Open Label)

Number of Arms

1

Arms and Interventions

Participant Group / Arm

Intervention / Treatment

Experimental: Patient Participants; This single-group study will recruit patients through the PI's clinical practice who are undergoing invasive neurophysiological monitoring (sEEG) with clinically necessary placement of electrodes in the supratemporal plane. All participants will complete the same behavioral response paradigms.

Behavioral: Dimension-Based Statistical Learning; Each participant will complete self-paced blocks of stimuli that will first establish a baseline for neural activity and behavioral responses with clear speech, and will then record responses for experimentally manipulated blocks to introduce 1) speech-in-noise and 2) a Canonical-Reverse block to model an "accent." Auditory stimuli will be adjusted to a comfortable level for each participant as determined by a calibration process completed by the participant. Each block involves listening to sound via earphones and making a categorical decision between initial consonants (/b/ or /p/) by tapping a button to indicate the word heard by the participant.

What is the study measuring?

Primary Outcome Measures

Outcome Measure	Measure Description	Time Frame
Supratemporal Neural Response to Change in Acoustic-Phonetic Dimensions	Neural activity will be measured via simultaneous EEG-sEEG monitoring in the supratemporal plane as indicated by high-gamma band activity in the electrical signal. Neural activity will be measured as participants listen to acoustic stimuli with gradually manipulated acoustic dimensions, fundamental frequency (F0) and voice onset time (VOT). The data reported here is the number of temporal lobe channels demonstrating significant encoding of change in acoustic dimension (F0 as VOT is held constant).	During sEEG-EEG recording sessions, up to 3 hours total
Behavioral Impact of Change in Acoustic-Phonetic Dimensions	Behavioral responses in the form of a category judgment will be obtained as participants listen to acoustic stimuli in with gradually varying fundamental frequency (F0) and voice onset time (VOT). Participants will provide a behavioral response by indicating the phoneme perceived at the beginning of stimulus words (/b/ or /p/). Specifically, the outcome is reported as percent of stimuli classified as /p/ over varying F0 with VOT held constant.	During sEEG-EEG recording sessions, up to 3 hours total
Supratemporal Neural Response to Change in Listening Context	Neural activity was measured via sEEG monitoring as indicated by high-gamma band activity in the electrical signal. Neural activity will be measured as participants listen to acoustic stimuli in accented speech. The un-transformed voltage represents the difference in electric potential between a specific electrode contact and the reference electrode contact; since we use a common average reference, that means it's the difference between a specific electrode contact and the mean voltage across all electrodes. We then z-score to characterize shifts from baseline activity (where z = 0) at a specific electrode, which is believed to measure changes in voltage due largely to post-synaptic currents. This is a mathematical transformation rather than a published scale or standardized assessment. More extreme z-scores (+ or -) indicate a greater change from baseline local neural activity; in other words, stimuli evoked greater activity in this region.	During sEEG-EEG recording sessions, up to 3 hours total
Behavioral Impact of Change in Listening Context	Behavioral responses in the form of a category judgment will be obtained as participants listen to acoustic stimuli in a varied listening context: accented speech. Participants will provide a behavioral response by indicating the phoneme perceived at the beginning of stimulus words (/b/ or /p/). Specifically, the outcome is reported as the mean % of stimuli classified as /p/ with varying VOT and F0 relationships.	During sEEG-EEG recording sessions, up to 3 hours total

Secondary Outcome Measures

Outcome Measure	Measure Description	Time Frame
Neural Response of Non-Regions of Interest to Change in Acoustic-Dimension	Neural activity will be measured via simultaneous EEG-sEEG monitoring in the supratemporal plane as indicated by high-gamma band activity in the electrical signal. Neural activity will be measured as participants listen to acoustic stimuli with gradually manipulated acoustic dimensions, fundamental frequency (F0) and voice onset time (VOT). The data reported here is the number of temporal lobe channels demonstrating significant encoding of change in acoustic dimension (F0 as VOT is held constant).	During sEEG-EEG recording sessions, up to 3 hours total
Neural Response of Non-Regions of Interest to Change in Listening Context	Neural activity was measured via sEEG monitoring as indicated by high-gamma band activity in the electrical signal. Neural activity will be measured as participants listen to acoustic stimuli in accented speech. The un-transformed voltage represents the difference in electric potential between a specific electrode contact and the reference electrode contact; since we use a common average reference, that means it's the difference between a specific electrode contact and the mean voltage across all electrodes. We then z-score to characterize shifts from baseline activity (where z = 0) at a specific electrode, which is believed to measure changes in voltage due largely to post-synaptic currents. This is a mathematical transformation rather than a published scale or standardized assessment. More extreme z-scores (+ or -) indicate a greater change from baseline local neural activity; in other words, stimuli evoked greater activity in this region.	During sEEG-EEG recording sessions, up to 3 hours total

Collaborators and Investigators

• Sponsor: University of Pittsburgh
• Collaborators: National Institutes of Health (NIH); National Institute on Deafness and Other Communication Disorders (NIDCD); Carnegie Mellon University
• Investigators: Principal Investigator: Taylor J Abel, MD, University of Pittsburgh

Study record dates

Study Major Dates

• Study Start (Actual): May 2, 2022
• Primary Completion (Actual): July 1, 2024
• Study Completion (Actual): April 10, 2025

Study Registration Dates

• First Submitted: January 12, 2022
• First Submitted That Met QC Criteria: January 12, 2022
• First Posted (Actual): January 26, 2022

Study Record Updates

• Last Update Posted (Estimated): September 8, 2025
• Last Update Submitted That Met QC Criteria: August 14, 2025
• Last Verified: August 1, 2025

More Information

Terms related to this study

• Keywords: sEEG; Speech representation; Supratemporal plane
• Additional Relevant MeSH Terms: Brain Diseases; Central Nervous System Diseases; Nervous System Diseases; Epilepsy
• Other Study ID Numbers: STUDY21090011; 1R21DC019217-01A1 (U.S. NIH Grant/Contract)

Plan for Individual participant data (IPD)

• Plan to Share Individual Participant Data (IPD)?: YES

IPD Plan Description

The individual deidentified participant data intended to be shared include the individual participant data that underlie the results to be reported in published articles after deidentification.

Other documents that will be made available include the study protocol and statistical analysis plan.

Data will be available as soon as possible following publication, but no later than one year upon completion. There is no end date.

IPD will be made available for any purpose via open access.

• IPD Sharing Time Frame: Data will be made available as soon as possible, but no later than one year upon completion of the study.
• IPD Sharing Access Criteria: Our data will be made publicly available online as soon as possible. Data will be easily and widely accessible.
• IPD Sharing Supporting Information Type: STUDY_PROTOCOL; SAP

Drug and device information, study documents

• Studies a U.S. FDA-regulated drug product: No
• Studies a U.S. FDA-regulated device product: No
• product manufactured in and exported from the U.S.: No