Feedforward and feedback in speech perception: Revisiting analysis by synthesis

Brain regions recruited for the effortful comprehension of noise-vocoded words

“…However, on the basis of the present data, we propose that the effortful perception of noise-vocoded speech, and perhaps other forms of degraded speech also, is assisted by networks involving regions of the dorsal auditory pathway that decode speech using nonacoustic, articulatory templates. This accords well with the suggestion made by , Iacoboni (2008), Poeppel and Monahan (2010) and others that these regions provide an internal simulation that helps to match degraded speech input to internal templates derived from a prototypical motor pattern. Just such a feedback process has been invoked to explain the perceptual learning of noise-vocoded speech (HervaisAdelman et al, 2008), and the current functional imaging study therefore provides initial activation evidence for brain regions that contribute to this form of perceptual feedback.…”

Sensorimotor adaptation affects perceptual compensation for coarticulation

“…One of the central questions in speech perception is whether and to what extent our perception of an acoustic speech signal maps onto the physical mechanisms utilized to produce that sound (Liberman and Mattingly, 1989). While some theories claim that representations accessed during speech perception can be described succinctly in terms of acoustics (e.g., Blumstein and Stevens, 1981), others posit that perception involves accessing representations more directly related to the articulatory gestures that generated the speech (e.g., Liberman and Mattingly, 1985;Fowler, 1986;Poeppel and Monahan, 2011). This study sought to contribute to this debate by asking whether altering speakers' articulation-to-sound mapping for the production of a vowel has consequences for their perception of coarticulated consonants whose interpretation is dependent on vowel context.…”

Prediction of Agreement and Phonetic Overlap Shape Sublexical Identification

“…Given the abundant noise and variation that exist in the physical speech signal, the ability of the perceptual system to integrate various knowledge sources would clearly be advantageous (for a review from a speech processing perspective, see Uslar et al, 2013;Wendt, Brand, & Kollmeier, 2014). Many models implement some form of top-down information flow as a core tenet, including interactive-activation models (Bowers & Davis, 2004;McClelland & Elman, 1986) and forward-models (Gagnepain, Henson, & Davis, 2012;Halle & Stevens, 1962;Martin, 2016;Poeppel & Monahan, 2011). Others interpret such "top-down" findings as ancillary task effects only affecting decision-stage nodes (Norris, McQueen, & Cutler, 2000).…”