Neurocognitive research on Parkinson’s disease (PD) points to action concept processing as a key target for patient identification and characterization. This domain hinges on electrophysiological mechanisms typically compromised in the disorder. Yet, very few studies have examined such mechanisms in PD, and none has done so integrating event-related potentials (ERPs) and time-frequency (TF) measures. Also, most paradigms require fast motor responses, proving blind to how patients process such concepts in the absence of stringent time constraints. To bridge these gaps, we assessed ERPs and TF modulations as healthy persons (HPs) and PD patients performed a delayed-response semantic judgment task targeting action concepts.

Processing of action concepts (semantic units denoting bodily movements) is vital for normal functionality, as they mediate everyday life interactions with objects and people (Amoruso et al., 2013, van Elk et al., 2009). Evidence from HPs indicates that action concepts hinge on both neural and peripheral motor systems. Specifically, their processing distinctly engages motor brain regions (García et al., 2019, Pulvermüller, 2013) and networks (Moguilner et al., 2021a, Moguilner et al., 2021b), is differentially affected by M1 neurostimulation (Birba et al., 2020b, Tomasino et al., 2008, Vukovic et al., 2017), and it can modulate the planning and execution of bodily actions (García and Ibáñez, 2016a, García and Ibáñez, 2016b, Mirabella et al., 2012). Since motor system disruptions are the core signature of PD (Bloem et al., 2021), action concepts have been proposed as a strategic target to capture early markers of the disorder –a core claim of recent translation models, such as the disrupted motor grounded hypothesis (Birba et al., 2017).

Indeed, PD patients are impaired in comprehending (Bocanegra et al., 2017a; Fernandino et al., 2013; García et al., 2018a), associating (Bocanegra et al., 2015), and verbally expressing (Cotelli et al., 2007, Piatt et al., 1999, Rodríguez-Ferreiro et al., 2009, Signorini and Volpato, 2006) action concepts –although some studies have reported preserved behavioral performance in specific tasks (Aiello et al., 2022, Møller et al., 2023). Importantly, these deficits hold irrespective of patients’ executive and domain-general dysfunctions, speaking to their systematicity beyond cognitive impairment (Bocanegra et al., 2015, García et al., 2017, García et al., 2018a). Moreover, they discriminate among disease subtypes (Bocanegra et al., 2017a; García et al., 2018a, García et al., 2022a), capture disease-specific neural disruptions (Abrevaya et al., 2017, Birba et al., 2022a, Gallese and Cuccio, 2018), and do not emerge in disorders sparing motor circuitry (Birba et al., 2022a, Moguilner et al., 2021a, Moguilner et al., 2021b). Briefly, together with exaggerated congruency effects, known as hyperpriming (Filoteo et al., 2003, Marí-Beffa et al., 2005, Spicer et al., 1994), action-concept anomalies constitute one of the best-established semantic particularities of PD.

However, most action concept research on PD exclusively employs behavioral tasks, failing to reveal relevant brain disruptions and precluding the integration of findings with neurobiological models of the disease. In particular, very few studies have used electroencephalography (EEG), a portable, low-cost technique that has revealed robust signatures of action concept processing in HPs (Amoruso et al., 2013, Cuellar and Del Toro, 2017). Relevant insights could be obtained by two EEG metrics: ERPs and TF modulations.

First, ERPs are time-locked voltage changes relative to a given event (e.g., stimulus presentation) (Sur and Sinha, 2009). In particular, the N400 component (a negative deflection peaking between 250 to 600 ms after stimulus presentation) (Kutas and Federmeier, 2011) is a key marker of the integration of action concepts evoked by images, videos, and verbs (Amoruso et al., 2013, Cervetto et al., 2021). In HPs, indeed, early (~250-300 ms) fronto-central N400 effects index the processing of action images (Amoruso et al., 2013) and efficient detection of incongruency between them (Amoruso et al., 2013, Bach et al., 2009, Mudrik et al., 2010, Proverbio and Riva, 2009, Proverbio et al., 2010), whereas later (~350-500 ms) effects are observed for action verbs (Cervetto et al., 2021, Dalla Volta et al., 2018, Zhou et al., 2022).

Second, TF modulations are ongoing changes in the synchronization of EEG signals over specific frequency bands (Morales and Bowers, 2022). In HPs, frontocentral theta (4-8 Hz) enhancement underlies semantic retrieval and incongruency detection (Maguire et al., 2015, Reid et al., 2009), driven by contextual expectations and conflict monitoring (Urgen et al., 2013). Also, theta modulations often precede the desynchronization of mu rhythms (8-14 Hz) (Urgen et al., 2013), commonly observed during action observation and execution (Coll et al., 2017, Pineda et al., 2000, Quandt et al., 2012, Quandt et al., 2013) –a hallmark signature of covert action simulation, traceable to the mirror neuron system (Debnath et al., 2019, Pineda, 2005). Similar effects emerge in the face of action images during verbal (Cuellar and Del Toro, 2017) and non-verbal (Pfurtscheller et al., 2006) tasks, including semantic judgment (Moreno et al., 2015).

Yet, to our knowledge, no previous study has jointly examined ERP and TF correlates of action semantic processing in PD. A recent study on the disease found abolished N400 modulations when action concepts occur in incongruent events, pointing to abnormal semantic integration (Wyrobnik et al., 2022). Also, though theta and mu oscillations remain unexplored in action-concept research on PD, they are attenuated over mid-frontal electrodes when patients face conflict resolution and error adjustment demands (Cavanagh et al., 2018, Singh et al., 2018) –as those involved in action relatedness tasks. Though incipient, then, action-concept EEG research on PD has potential to uncover fine-grained markers of the disorder.

Here, HPs and PD patients performed an action relatedness judgment task featuring related and unrelated picture pairs and requiring delayed responses (a strategic approach to circumvent speed-related errors in PD). In each group, we first tested for action relatedness effects, namely, the response time (RT) difference between unrelated and related trials. We then established task-specific N400 and theta/mu modulations in HPs, replicated the analysis on PD patients, and examined whether detected effects were associated with participants’ cognitive outcomes.

We raised three sets of hypotheses. First, we predicted that, relative to related trials, unrelated trials in HPs would yield slower responses, increased early N400 modulations, higher theta enhancement, and reduced power decreases in mu rhythms. Second, we anticipated that PD patients would also exhibit a behavioral relatedness effect (since the task gives ample time to respond), but that early electrophysiological effects observed in HPs would be abolished or altered. Finally, in light of previous works, we hypothesized that semantic EEG indexes would not correlate with patients’ cognitive outcomes. By testing these conjectures, we aim to illuminate the neurocognitive bases of action semantic processing in PD.