This symposium brings together leading researchers to examine how the brain anticipates forthcoming speech—from prosodic cues and lexical forms to semantic content. Topics include neural markers of prediction in the sub-lexical level, anticipatory brain activity in sentence comprehension, the role of sensorimotor systems in semantic processing, and predictive mechanisms during naturalistic language understanding. This event provides an in-depth overview of recent advances in the study of predictive processing in speech.

Program

13.10   Opening

13.15   The Pre-Activation Negativity, PrAN (Mikael Roll)

13.45   Anticipatory negativities in sentence comprehension: the role of word certainty (Patricia León-Cabrera)

14.15   Coffee break

14.45   Brain correlates of perceptual, semantic, and action prediction (Luigi Grisoni)

15.15   Predictive processing during naturalistic language comprehension: Evidence from neuroimaging and eye movements (Floris de Lange)

15.45   Break

15.55   Panel discussion

** If you are joining in person, please sign up here so we can prepare the fika.

Abstracts

1. The Pre-Activation Negativity (PrAN)

Mikael Roll (Lund University)

Roll et al. (2009, 2010) reported a P2 enhancement in response to salient speech melodies. However, Söderström et al. (2012) found that less salient melodies provided stronger cues for linguistic prediction. Building on this, Roll et al. (2015) proposed that the observed effect was not a P2 enhancement, but rather a negative potential associated with predictive utility, which they termed the pre-activation negativity (PrAN). This interpretation was supported by findings that PrAN amplitude inversely correlates with the number of lexical competitors at word onset (Söderström et al., 2016). More recent studies have confirmed PrAN as a forward-looking component of linguistic processing (Hjortdal et al., 2024), distinct from and co-occurring with the N400 (Kwon & Roll, 2024).

2. Anticipatory negativities in sentence comprehension: the role of word certainty

Patricia León-Cabrera (Universitat Internacional de Catalunya)

Slow and sustained event-related negativities have long been a hallmark of anticipation in a variety of cognitive domains. In recent years, different bodies of work have converged in reporting negative-going potentials preceding words or word segments that can be pre-activated based on linguistic cues (reviewed in León-Cabrera et al., 2024). Specifically, our work has identified two negativities when sentence-final words are strongly expected in a context (“ball” after “The goalkeeper managed to catch the”): a slow widespread negativity that builds up during sentence processing, and a later, left anterior negativity that immediately precedes the sentence-final word. In this talk, I will summarize this evidence and present a study examining the influence of sentence-final word certainty on the emergence of these anticipatory negativities. I will discuss how these slow negativities, combined with established components, can help disentangle prediction-related processes occurring at different stages and timescales during language comprehension.

3. Brain correlates of perceptual, semantic, and action prediction

Luigi Grisoni (Brain Language Laboratory, Department of Philosophy and Humanities, Freie Universität Berlin; Cluster of Excellence ‘Matters of Activity. Image Space Material’, Humboldt Universität zu Berlin)

Semantic and conceptual grounding theories emphasize functional interaction between perceptual, motor, and conceptual knowledge (Pulvermüller, 1999; Barsalou, 2008). Although much evidence shows consistent activation in sensorimotor brain areas during language comprehension (e.g., Hauk et al., 2004; Grisoni et al., 2016), it has been argued that such activations may index post-hoc, epiphenomenal re-processing (Mahon and Caramazza, 2008; Bedny and Caramazza, 2011) rather than genuine semantic processing. In recent works, however, a significant contribution of sensorimotor areas has been reported before predictable words appear in comprehension (Grisoni et al., 2017; Grisoni et al., 2021; León-Cabrera et al., 2024) and before predictable words are pronounced to complete sentence fragments (Grisoni et al., 2024). Sentence fragments that strongly predict subsequent words induced slow-wave potentials before the expected words; this potential was weaker if the preceding fragments were unpredictable. That this anticipatory activity indexed predictive semantic processing was further demonstrated by the observation of cortical sources in specific sensorimotor brain areas for action-related (e.g., action verbs, tool nouns) words but in posterior, visual, areas for visual-related words (e.g., animal nouns). Furthermore, inverse correlations between the predictive signal and the well-known brain index of semantic processing, N400, suggest that these two responses have a similar semantic discriminatory function. Overall, these data show that activity in sensorimotor brain areas also reflects the meaning subjects expect and, therefore, it cannot originate from re-processing but from genuine semantic processing.

4. Predictive processing during naturalistic language comprehension: Evidence from neuroimaging and eye movements

Floris de Lange (Donders Institute, Radboud University Nijmegen), on Zoom

In my talk, I will discuss neuroimaging and eye-tracking evidence from participants listening to audiobooks, and participants reading books,  suggesting that the brain engages in probabilistic hierarchical prediction of future input. 

Hosted by Neurolinguistics in Sweden (NLS).