Breaking the mirror: Asymmetrical disconnection between the phonological input and output codes

Charlotte Jacquemot, Emmanuel Dupoux, Anne‐Catherine Bachoud‐LéviCognitive 2006 · DOI

TL;DR

This paper investigates the relationship between speech perception and production systems by studying a patient with conduction aphasia. They found that the patient had a specific deficit in transferring phonological information from perception to production, suggesting these systems are separate but linked by distinct conversion mechanisms that can be independently damaged.

What Problem Does This Paper Address?

The paper tackles a fundamental question in psycholinguistics: how are the systems responsible for understanding spoken language (perception) and producing spoken language (production) organized? Specifically, it explores whether the phonological representations used in these two systems are shared or separate, and how they interact. Previous research offered conflicting views: some models proposed a single, shared phonological code, while others suggested separate input and output codes. Understanding this architecture is crucial for explaining language disorders like aphasia, where disruptions in these systems can lead to specific communication difficulties. The authors aim to clarify this by examining a patient with conduction aphasia, a condition known for impaired repetition.

Methods

The researchers employed a case study approach, focusing on a single patient, F.A., who had conduction aphasia. They designed a series of tasks to probe different aspects of speech processing:

  • Speech Perception Tasks:
    • AX Discrimination Task: Assessed phonological decoding by presenting pairs of sounds and asking the patient to judge if they were the same or different, even when spoken by different speakers.
  • Auditory Word-Picture Matching Task: Evaluated semantic comprehension by presenting an auditory word and a picture, asking the patient to determine if the word named the picture. This task included various distractor types (unrelated, semantic, phonological) to pinpoint the nature of any deficit.
  • Speech Production Tasks:
    • Picture Naming Task: Assessed general speech production by asking the patient to name pictures.
  • Word and Pseudoword Repetition Task: This was a key task, designed to specifically test the link between phonological input and output. The patient repeated real words and non-words (pseudowords) of varying lengths and frequencies. The differential performance on words versus pseudowords was critical for their hypothesis.
  • Reading Task: To rule out a general production deficit, the patient's ability to read words and pseudowords was tested.
  • Rhyme Judgement Task (written words): This task aimed to assess the transfer of phonological information from the output (reading a word and converting it to sound) to the input (comparing sounds for rhyme).
  • Auditory Item-Picture Matching Task: This task further probed the production-to-perception link by presenting a picture and an auditory item (word or pseudoword) and asking the patient to judge if the auditory item rhymed with or was contained within the picture's name, without speaking.

Datasets

The study utilized several sets of stimuli:

  • CV-CVC bisyllables: Used in the AX discrimination task, these were minimal pairs differing by a single phonetic feature.
  • Common concrete nouns: Selected for the auditory word-picture matching task.
  • Words and Pseudowords: Used extensively in repetition, reading, and auditory item-picture matching tasks. These were carefully controlled for length and frequency.
  • Picture stimuli: Used for naming and auditory item-picture matching tasks.

No new benchmarks were introduced; rather, existing tasks were adapted and applied to the patient.

Key Insights and Ideas

  • Asymmetrical Disconnection: The most striking finding is the asymmetrical nature of the deficit. F.A. struggled significantly with repeating pseudowords but performed relatively well with real words, and her speech perception was largely intact. This points to a specific problem in transferring phonological information from the perceptual input system to the production output system, rather than a general deficit in either perception or production.
  • Separate but Linked Phonological Codes: The results strongly support a model where phonological processing in perception and production involves distinct, but connected, codes. The pseudoword repetition deficit suggests that the non-lexical route, crucial for processing novel sound sequences, is impaired.
  • Independent Conversion Mechanisms: The study provides evidence for two distinct conversion mechanisms: one from phonological input to output (impaired in F.A.) and another from phonological output to input (intact in F.A.). This implies that these pathways can be damaged independently, offering a more nuanced view of language processing architecture.
  • Implications for Models of Aphasia: The findings help refine models of conduction aphasia, suggesting that the deficit may not solely stem from a phonological impairment in perception or production, but rather from a disconnection between these systems at the phonological level.

Key Results

  • Speech Perception: F.A. performed at near-normal levels on phonological decoding (AX discrimination) and semantic comprehension (word-picture matching) tasks, indicating intact input phonology and lexical access.
  • Speech Production: While F.A. showed slight impairments in picture naming and word reading, her performance was significantly worse in pseudoword repetition (35.4% correct) compared to word repetition (82.3% correct). This difference was qualitative as well, with different error patterns.
  • Production-to-Perception Link: Tasks designed to assess the transfer of phonological information from output to input (rhyme judgment, auditory item-picture matching) showed that F.A.'s performance, while not perfect, was comparable to her reading performance and did not differ significantly between word and pseudoword conditions. This suggests the output-to-input conversion mechanism was intact.

Limitations and Future Work

  • Case Study: The findings are based on a single patient, which limits generalizability. While F.A.'s profile is informative, further studies with similar patients are needed to confirm these results.
  • Anatomical Correlation: The study notes that the arcuate fasciculus, a potential anatomical substrate for the proposed bidirectional connections, was not clearly visible on F.A.'s lesion scans, preventing direct anatomical confirmation of the lesion's impact on this pathway.
  • Future Research: The authors suggest that future research should focus on testing predictions regarding bidirectional connections within the arcuate fasciculus and further investigating how these conversion mechanisms can be selectively damaged. They also propose that these findings have implications for understanding other language disorders and for refining computational models of speech processing.

Summary updated

Abstract

In this paper, we study the link between the processing systems that sustain speech perception and production in a patient (F.A.) with conduction aphasia. Her pattern of performance in repetition task - quantitative but also qualitative striking difference in errors with pseudowords versus words - cannot be properly accounted for either by a perception deficit or by a production deficit. We discuss this finding according to theoretical models of phonological processing and show that it is best explained by an impaired ability to transfer phonological information from the perception to the production system. We also probed for a phonological link in the opposite direction, from the production to the perception system. F.A.'s results show that this link was not impaired. Overall, our results suggest that (a) the phonological codes in perception and in production are separate but connected by two conversion mechanisms and that (b) these two mechanisms can be disrupted independently.
BibTeX
@inproceedings{jacquemot2006breaking,
  author = {Charlotte Jacquemot and Emmanuel Dupoux and Anne‐Catherine Bachoud‐Lévi},
  title = {{Breaking the mirror: Asymmetrical disconnection between the phonological input and output codes}},
  booktitle = {Cognitive},
  year = {2006},
  doi = {10.1080/02643290600683342},
}
Figure