◐ Shell
reader mode source ↗
Skip to main page content
Advanced
User Guide
Full text links Cite
Display options

Abstract

The auditory cortex communicates with the frontal lobe via the middle temporal gyrus (auditory ventral stream; AVS) or the inferior parietal lobule (auditory dorsal stream; ADS). Whereas the AVS is ascribed only with sound recognition, the ADS is ascribed with sound localization, voice detection, prosodic perception/production, lip-speech integration, phoneme discrimination, articulation, repetition, phonological long-term memory and working memory. Previously, I interpreted the juxtaposition of sound localization, voice detection, audio-visual integration and prosodic analysis, as evidence that the behavioral precursor to human speech is the exchange of contact calls in non-human primates. Herein, I interpret the remaining ADS functions as evidence of additional stages in language evolution. According to this model, the role of the ADS in vocal control enabled early Homo (Hominans) to name objects using monosyllabic calls, and allowed children to learn their parents' calls by imitating their lip movements. Initially, the calls were forgotten quickly but gradually were remembered for longer periods. Once the representations of the calls became permanent, mimicry was limited to infancy, and older individuals encoded in the ADS a lexicon for the names of objects (phonological lexicon). Consequently, sound recognition in the AVS was sufficient for activating the phonological representations in the ADS and mimicry became independent of lip-reading. Later, by developing inhibitory connections between acoustic-syllabic representations in the AVS and phonological representations of subsequent syllables in the ADS, Hominans became capable of concatenating the monosyllabic calls for repeating polysyllabic words (i.e., developed working memory). Finally, due to strengthening of connections between phonological representations in the ADS, Hominans became capable of encoding several syllables as a single representation (chunking). Consequently, Hominans began vocalizing and mimicking/rehearsing lists of words (sentences).

Keywords: aphasia; auditory dorsal stream; auditory ventral stream; evolution; language; speech.

PubMed Disclaimer

Figures

Figure 1
Figure 1
The neuroanatomy of the auditory ventral and dorsal streams. Two pathways connect the auditory cortex and the inferior frontal gyrus (IFG): the auditory ventral stream (AVS; red arrows), which processes sound recognition, and the auditory dorsal stream (ADS; blue arrows), which processes sound localization, speech production and repetition, phonological working memory, phonological long-term memory and more. In the AVS, the anterior superior temporal gyrus (aSTG) communicates with the IFG via relay stations in the middle temporal gyrus (MTG) and temporal pole (TP). In the ADS, the posterior superior temporal gyrus (pSTG) communicates with the IFG via relay stations in the posterior superior temporal sulcus (pSTS), Sylvian parietal-temporal junction (Spt) and inferior parietal lobule (IPL).
Figure 2
Figure 2
Stage 1: Mimicry and the imitation of lip-movements. The model proposes that early Hominans evolved to mimic vocalizations by reading lips to learn novel words. Here, an adult teaches a child the vocalization associated with a rabbit (Left). The adult recognizes the rabbit through auditory object recognition in the aSTG (A) or through visual object recognition in the inferior temporal gyrus (ITG; B) and associates it with the concept of a rabbit that is coded in the semantic lexicon of the MTG-TP (C). The adult vocalizes the call associated with a rabbit by projecting from the semantic representation in the MTG to the praxic representations in area Spt and then to the IFG (E). The child learns the call associated with the animal by repeating the call. When the child hears the call, he/she encodes the acoustic properties of the call in the sound recognition center of the aSTG (A) and learns to associate the call with its semantic representation (C). In parallel, the pSTG-pSTS receive their own afferents from the auditory cortex and extract phonemic information from it. Via processing in the pSTG-pSTS of the ADS, the child then integrates the phoneme with its corresponding lip movements (D). The pSTG-pSTS then activate the praxic representation of the call in area Spt and then in the IFG (E). Finally, via feedback connections from the IFG to the Spt-pSTG, the child recognizes the emitted call as self-produced and verifies that the emitted call is acoustically similar to the call that he/she previously perceived (F). The child repeats this process until he/she can vocalize the call associated with a rabbit on his/her own.
Figure 3
Figure 3
Stage 2: Vocal mimicry at infancy and the liberation from imitating lip-movements. Top: As the infant mimicked each of the parents' calls, the acoustic properties of the call were encoded in the aSTG (A). In parallel, the phonemic and visemic representations of the call (B) were integrated with its praxic representation in area Spt (C). The infant then produced the call by projecting to the motor regions of the IFG (D). The long-term memory store for the vocal properties of calls in the Spt-IPL region is called the phonological lexicon. As during mimicry, sound recognition and the integration of phonemes with their visemes occur at the same time, associations formed between the acoustic representations of the aSTG and the corresponding phonological representations of the ADS (arrows between A and B). Middle: Individuals, who passed the vocal mimicry stage during infancy, also recognized sounds via the aSTG (A) and extracted its meaning via the MTG-TP (B). Given the presence of phonemic-visemic representations in pSTG that are associated with the perceived sound, individuals were capable of mimicking the call via aSTG to pSTG connections (arrow between A and C). Moreover, given the presence of phonemic-visemic-praxic (phonological) representations in the Spt-IPL region, individuals were capable of naming the object they see/hear via MTG-TP to Spt-IPL connections (arrow between B and D). Activation of the phonological representation then activated the motor cortex in the IFG, which initiated the vocalization of the word (E). Bottom: After years of practicing the word without relying on lip movement imitation, the connections with the visemic representations become less robust, and the phonological representations eventually came to comprise primarily the phonemic and praxic representations.
Figure 4
Figure 4
Computational views of the auditory ventral and dorsal streams. Top: In the ADS, repetition of a word (e.g., “UNIVERSITY”) is accomplished in two stages. While hearing the word (top), each syllable is encoded in order into a temporary memory buffer. Only after all syllables have been encoded (bottom), are the syllables extracted from the storage space (in the same order) to vocally produce the word. Bottom: In the AVS, a word (e.g., “ANTELOPE”) is recognized while being perceived. The first utterance (“A”) activates many possible matches, and the perception of each additional utterance further narrows the number of possibilities until only one match remains.
Figure 5
Figure 5
Stage 3: The poly-syllabic lexicon and the enhancement of working memory. Left: The model proposes that the transition from monosyllabic to polysyllabic repetition of calls was made possible by the development of a storage capacity in working memory. Speech repetition is initiated when the aSTG of the AVS perceives a call. As the aSTG-MTG extracts the call's meaning (not shown in figure), the aSTG also recognizes the individual syllables in the order in which they were perceived (A). Each acoustic syllabic representation in the AVS activates its corresponding monosyllabic phonological representation in the pSTG-Spt-IPL regions of the ADS (B). The Spt-IPL then activates the corresponding praxic representation in the IFG (C). The storage capacity in working memory was made possible by the development of inhibitory connections in which each monosyllabic acoustic representation in the aSTG suppresses the phonological representation of the syllable that succeeds it in the pSTG-Spt-IPL (T-shaped arrows). Because of this process, the succeeding syllable is dis-inhibited and vocalized only after the present syllable has been vocalized.
Figure 6
Figure 6
Stage 4: Chunking and the emergence of sentences. Top: The model proposes that, with the advent of chunking, constant rehearsal of a polysyllabic word (left) resulted in the word being encoded in the phonological lexicon as a single representation (right). Bottom: The encoding of a word as a single representation allowed individuals to vocalize and repeat lists of words. These lists could then be used to teach a sequence of actions. For example, the figure shows an adult teaching a child how to spear a fish and cook it over a fire by vocalizing 3 words: fish – spear – fire. The child can then repeat and thus memorize the sequence, and later he will be able to perform the sequence of actions on his own.
See this image and copyright information in PMC

References

    1. Acheson D. J., Hamidi M., Binder J. R., Postle B. R. (2011). A common neural substrate for language production and verbal working memory. J. Cogn. Neurosci. 23, 1358–1367. 10.1162/jocn.2010.21519 - DOI - PMC - PubMed
    1. Acheson D. J., MacDonald M. C. (2009). Twisting tongues and memories: Explorations of the relationship between language production and verbal working memory. J. Mem. Lang. 60, 329–350. 10.1016/j.jml.2008.12.002 - DOI - PMC - PubMed
    1. Ahveninen J., Jaaskelainen I. P., Raij T., Bonmassar G., Devore S., Hämäläinen M., et al. (2006). Task-modulated “what” and “where” pathways in human auditory cortex. Proc. Natl. Acad. Sci. U.S.A. 103, 14608–14613. 10.1073/pnas.0510480103 - DOI - PMC - PubMed
    1. Anderson J. M., Gilmore R., Roper S., Crosson B., Bauer R. M., Nadeau S., et al. (1999). Conduction aphasia and the arcuate fasciculus: a reexamination of the Wernicke-Geschwind model. Brain Lang. 70, 1–12. 10.1006/brln.1999.2135 - DOI - PubMed
    1. Andics A., Gácsi M., Faragó T., Kis A., Miklósi Á. (2014). Voice-sensitive regions in the dog and human brain are revealed by comparative fMRI. Curr. Biol. 24, 574–578. 10.1016/j.cub.2014.01.058 - DOI - PubMed
Show all 199 references

LinkOut - more resources