Lecture 18: Cognitive models and the neural basis of Speech Perception

Question 1

What are the key proposals of the motor theory of speech perception?

Answer 1

• Speech perception focuses on intended vocal gestures rather than acoustic signals.
• It uses a specialized module distinct from non-speech sound processing.

Reference: Liberman et al. (1967); Liberman & Mattingley (1985).

Question 2

How does categorical perception support the motor theory?

Answer 2

Speech sounds (e.g., /p/ in “pin” vs. “spin”) are perceived categorically despite acoustic variability.

Question 3

What evidence supports motor involvement in speech perception?

Answer 3

• Wilson et al. (2004): fMRI studies showed motor and premotor area activation during passive listening.

Methodology: Functional MRI scanning during speech perception tasks.

• Meister et al. (2007): TMS over premotor areas impaired phoneme discrimination in noise but not color discrimination.

Methodology: TMS studies testing motor area involvement in speech perception.

Question 4

What evidence challenges the motor theory?

Answer 4

• Burns & Ward (1978): Categorical perception found for musical intervals, not unique to speech.
• Kuhl & Miller (1978): Chinchillas showed similar phoneme boundary perception.

Methodology: Behavioral experiments testing animals’ responses to speech continua.

Question 5

What are the key components of the classic model of speech perception?

Answer 5

• Wernicke’s Area (Superior Temporal Gyrus): Responsible for speech perception.
• Broca’s Area (Inferior Frontal Gyrus): Responsible for speech production.

Left hemisphere dominance based on aphasia studies.

Question 6

What are the two streams in the dual-streams model?

Answer 6

• Ventral Stream: Processes word recognition and meaning (“What?”).

Bilateral involvement.

• Dorsal Stream: Links perception to production, aids speech learning (“How?”).

Left hemisphere dominance.

Reference: Hickok & Poeppel (2007).

Question 7

What evidence supports the dual-streams model?

Answer 7

• Ventral stream: Damage to the anterior temporal lobe causes semantic impairments (Bates et al., 2003).
• Dorsal stream: Listening to syllables activates motor areas (Wilson et al., 2004).

Methodology: Functional MRI and behavioral studies.

Question 8

What is the cohort model, and how does it work?

Answer 8

• Words are activated immediately upon minimal input, with multiple candidates competing until the uniqueness point is reached.

Reference: Marslen-Wilson & Tyler (1981).

Question 9

What evidence supports the cohort model?

Answer 9

• Shadowing Task: Listeners recognized words before their full length was heard (250 ms vs. 375 ms).

Methodology: Behavioral task measuring response latency during real-time word recognition.

Question 10

What are the limitations of the cohort model?

Answer 10

Lacks computational implementation, making it less precise for predicting interactions.

Question 11

What are the key features of the TRACE model?

Answer 11

• Interactive activation with three levels: Acoustic features, phonemes, and words.
• Bidirectional excitatory connections (bottom-up and top-down feedback).
• Inhibitory competition within layers (e.g., lexical competition).

Reference: McClelland & Elman (1986).

Question 12

How does the TRACE model explain context effects?

Answer 12

• Top-down feedback allows word-level activity to influence phoneme perception.

Example: The Ganong Effect, where ambiguous /g/ is perceived as “gift” due to lexical context.

Question 13

What evidence supports the TRACE model?

Answer 13

• Eyetracking Studies: Eye movements reflected TRACE predictions during auditory word recognition tasks.

Study: Allopenna, Magnuson, & Tanenhaus (1998).
Methodology: Real-time tracking of gaze as participants heard and identified spoken words.

Question 14

How does context influence speech perception?

Answer 14

• Visual Context: The McGurk Effect shows how conflicting audio and visual cues produce altered perceptions (e.g., /ba/ + /ga/ = /da/).

Study: McGurk & MacDonald (1976).
Methodology: Behavioral experiments using mismatched audio and lip movements.

• Lexical Context: The Ganong Effect demonstrates how word knowledge biases phoneme perception.

Study: Ganong (1980).
Methodology: Testing ambiguous phoneme boundaries with lexical influence.

Question 15

What are the key insights from the motor theory?

Answer 15

Motor areas contribute to speech perception, but evidence challenges the claim that speech is processed by a unique module.

Question 16

How do the dual streams explain speech perception?

Answer 16

• Ventral Stream: Recognizes word meaning.
• Dorsal Stream: Links speech perception to production and motor functions.

Question 17

What are the strengths of the TRACE model?

Answer 17

Explains top-down effects (e.g., lexical bias) and real-time word recognition dynamics.

Question 18

How do cognitive models like the cohort and TRACE models differ?

Answer 18

• Cohort Model: Sequential word activation with a focus on early input and uniqueness points.
• TRACE Model: Parallel activation with interactive feedback across acoustic, phoneme, and word levels.

Question 19

Why is the Ganong Effect significant?

Answer 19

Demonstrates how higher-level lexical context shapes lower-level phoneme perception, supporting interactive models like TRACE.

Question 20

What limitations exist in traditional speech perception models?

Answer 20

They often fail to account for non-human evidence of categorical perception (e.g., in animals) and non-speech sound processing.