Parkinson’s disease (PD) is a progressive multisystem neurodegenerative disorder (Dickson, 2018) characterized at the clinical level by a constellation of motor and non-motor symptoms (Trojano and Papagno, 2018), among which cognitive impairments have received particular attention due to their consequences in everyday functioning [e.g., Leroi et al. (2012)]. Alongside the well-known deficits in executive functioning, visuo-spatial abilities and memory, recent evidence has underlined possible socio-cognitive impairments in PD. Social cognition is a complex cognitive domain, which refers to a set of different processes aimed at recognizing and interpreting signals from the environment, understanding self and others’ behaviors, and adapting the response based on social needs (Frith, 2008). As a multi-faceted domain, social processes required for successful social interaction include aspects of social perception (e.g., emotion recognition), theory of mind (ToM, also defined as mental states attribution), empathy and social behavior (Henry et al., 2016). Notwithstanding some controversial results, deficits of social cognition in PD have been previously reported, with the majority of studies showing significant emotion recognition deficits, particularly for negative emotions (Gray and Tickle-Degnen, 2010; Argaud et al., 2018; Coundouris et al., 2019). Interestingly, these deficits seem to be at least partially independent from depressive symptomatology (Gray and Tickle-Degnen, 2010; De Risi et al., 2018), executive dysfunctions (Saenz et al., 2013; Enrici et al., 2015) or early visual processing deficits (Mattavelli et al., 2021). From a neuroanatomical perspective, emotion recognition disorders have been previously related to neurodegenerative alterations in regions belonging to the mesocorticolimbic pathway (Ibarretxe-Bilbao et al., 2009; Baggio et al., 2012), including the amygdala (Diederich et al., 2016). Considering its role as a hub within neural networks responsible for multisensorial and affective processing, damage to this brain region is hypothesized to hamper not only emotion recognition in PD, but also the attribution of affective mental states to others (i.e., affective ToM) (Premack and Woodruff, 1978; Call and Tomasello, 2008; Shamay-Tsoory et al., 2010; Diederich et al., 2016). Consistently, ToM dysfunctions have been recently described in PD patients (Bora et al., 2015), although the extent to which these deficits involve cognitive (i.e., ability to infer other intentions or beliefs) or affective (i.e., ability to infer other emotions) subcomponents is still an open issue, with some evidence supporting an early impairment in the cognitive sub-component and a later involvement of affective processing [e.g., Poletti and Bonuccelli (2012)], and other supporting early alterations in both ToM aspects [e.g., Santangelo et al. (2012)]. Besides, to the best of our knowledge, only a few studies investigated both emotion recognition and affective ToM in PD (Enrici et al., 2015; Foley et al., 2019; Alonso-Recio et al., 2021), and the relationship between these two facets, as well as the presence of these deficits in the earliest disease stages, is still unclear.

Despite the above-mentioned evidence of socio-cognitive deficits in PD, social cognition is not included among the cognitive domains clinically assessed for the definition of PD cognitive status according to the 2012 Movement Disorder Society (MDS) criteria for the detection of mild cognitive impairment (MCI) (Litvan et al., 2012). These criteria include a two-level operational schema depending on the neuropsychological assessment. Together with an abbreviated evaluation (i.e., MDS Level I criteria), specific guidelines for a comprehensive cognitive assessment (i.e., MDS Level II criteria) were defined. According to these recommendations, a patient is classified as MCI when showing impairments on at least two neuropsychological tests (i.e., one impaired test in two different cognitive domains or two impaired tests in one cognitive domain) in specific cognitive domains (i.e., attention and working memory, executive function, language, memory, and visuo-spatial functions). A recent study (Czernecki et al., 2021) evaluated the prevalence of socio-cognitive deficits in PD patients characterized by the MDS level I criteria. Notably, this study showed socio-cognitive dysfunctions in 30% of the sample, of which 20% was classified as cognitively unimpaired (CU). However, the narrow neuropsychological assessment, including only a measure of cognitive screening (i.e., MoCA) and a global measure of executive functioning (i.e., Frontal Behavioral Inventory) prevented a full MCI characterization (Litvan et al., 2012). In this study, we aim to fill this gap by investigating socio-cognitive deficits in PD patients characterized according to MDS Level II MCI criteria (Litvan et al., 2012). In order to analyse the prevalence of socio-cognitive dysfunctions in the clinical setting, we focused on tests validated for the Italian population assessing social perception (emotion recognition) and ToM.

Increasing evidence reports significant socio-cognitive dysfunctions in PD. However, the extent of these deficits according to the disease stage, as well as their clinical relevance, is still unclear. Thus, in this study, we adopted two social tasks clinically standardized for the Italian population in order to assess emotion recognition and ToM in PD, with the final aim of investigating the prevalence of socio-cognitive deficits in a sample of non-demented patients classified according to MDS Level II criteria for MCI (Litvan et al., 2012).

Emotion recognition and ToM represent two core components of social cognition, driving social interaction through automatic and voluntary processes (Coricelli, 2005). Notably, these cognitive functions require the integrity of a set of specific and shared brain regions belonging to frontal and mesocorticolimbic circuits (Mier et al., 2010; Abu-Akel and Shamay-Tsoory, 2011), which can be significantly affected by PD neurodegenerative processes (Baggio et al., 2012; Diederich et al., 2016). In support of this evidence, we found significant deficits in both facial emotion recognition and mental states attribution in PD-MCI. Notably, despite previous authors suggested a progressive impairment in cognitive ToM and a later involvement of affective ToM with disease progression [e.g., Poletti and Bonuccelli (2012)], our results support the presence of early alterations in both cognitive and affective facets of mental states attribution (Santangelo et al., 2012; Enrici et al., 2015; Coundouris et al., 2020). This evidence is further supported by the results on PD-CU group, which showed an isolated deficit in SET-IA and SET-EA.

Together with mentalizing deficits, PD-MCI showed a significant impairment in negative emotion recognition in agreement with previous quantitative and qualitative literature reviews (Gray and Tickle-Degnen, 2010; Argaud et al., 2018). Notably, fear represented the most difficult emotion to recognize, confirming previous evidence [e.g., Mattavelli et al. (2021)], followed by anger and sadness. On the other hand, no significant deficit has been found in disgust recognition. This result is inconsistent with earlier studies reporting a disproportionate deficit in disgust recognition in PD [e.g., Suzuki et al. (2006)] ascribable to the disruption of the basal ganglia–insula system (Obeso et al., 2008) involved in the recognition of this emotion (Phan et al., 2002; Fusar-Poli et al., 2009). However, more recent quantitative approaches showed heterogeneous results when considering single emotions, and a deficit in disgust recognition was found only in 47% of the studies taken into consideration (Argaud et al., 2018). Many potential confounding factors, such as disease severity, medication, or mood disorders (Gray and Tickle-Degnen, 2010) could contribute to emotion recognition deficit in PD, causing high variability in study results. Despite we did not find a significant relationship in our sample with dopaminergic treatment or mood disorders, future studies should be devoted to fully elucidate the role of these factors in PD emotion recognition deficits.

Meta-analytical evidence concurs however in reporting a major impairment in the recognition of negative emotions rather than of positive ones, and this deficit has been previously associated with amygdalar (Diederich et al., 2016) and mesocorticolimbic alterations (Ibarretxe-Bilbao et al., 2009; Baggio et al., 2012) in PD. In this sense, it is interesting to underline that in the current study we found a specific correlation between emotion recognition abilities and affective ToM, evaluated through SET-EA. Considering that a low performance at this sub-task has been previously related to amygdalar structural damage in other neurological populations (Cerami et al., 2014; Campanella et al., 2021), these results suggest possible common underlying pathological mechanisms affecting both emotion recognition and attribution in these patients. This perspective opens new relevant research questions that should be further explored. Indeed, although recent models have underlined the role of these socio-cognitive facets in social interaction [e.g., Cassel et al. (2019)], it is still to be fully elucidated the role of socio-cognitive deficits (in terms of emotion recognition or ToM) in altering social behavior.

Despite the significant results at group level, when evaluating socio-cognitive performance in single subjects according to normative data we found a limited number of patients showing a deficitary performance. This possibly suggests in the PD group the presence of subtle alterations in socio-cognitive tasks, still below the threshold of clinical relevance. This hypothesis is further supported by the presence of patients showing a borderline performance in both emotion recognition and mental states attribution. When considering normative cut-off scores, 20% of patients showed a significant clinical deficit in global emotion recognition or mental state processing. This percentage is similar to what has been previously reported (i.e., 30%) (Czernecki et al., 2021). The mismatch could be possibly explained in light of the different criteria adopted to define MCI [i.e., MDS Level II criteria in the current study and MDS Level I criteria in Czernecki et al. (2021)]. Indeed, MDS Level I criteria require an abbreviated evaluation compared to Level II criteria, with a foreseeable effect on diagnostic certainty, extent of clinical characterization and MCI detectability (Litvan et al., 2012). In accordance with this consideration, we found a higher percentage of patients characterized by MCI (i.e., 29%) compared to Czernecki’s study (i.e., 15.6%) (Czernecki et al., 2021), and, consistently, a lower percentage of PD-CU (i.e., 11%) showing socio-cognitive deficits.

When we included social cognition among the cognitive domains for MCI criteria, two PD-CU patients were re-classified as PD-MCI multiple domains. Considering the contribution of MCI classification in predicting the hazard of PD dementia (Hoogland et al., 2017), this result indicates a possible benefit in considering social cognition among the MDS cognitive domains in order to improve MCI detectability. Consistently with previous literature findings (Litvan et al., 2012; Barvas et al., 2021), the classification of our PD sample according to MDS Level II criteria showed predominant executive functioning, memory, and visuospatial deficits, but none of the patients satisfied the criteria for PD-MCI single domain. This result is in line with previous evidence showing a prevalence of multiple domain impairments (Goldman et al., 2015), but suggests potential challenges in identifying domain-specific PD-MCI subtypes using MDS criteria.

The main limitations of this study are represented by the small sample size that might affect the statistical power of the analyses, and by the lack of a detailed neuropsychological characterization of healthy controls, which does not allow to exclude the presence in this sample of subtle deficits in single cognitive domains. Despite the MoCA cut-off score used to include healthy participants is lower than the one suggested by international meta-analytic results [e.g., Carson et al. (2018)], this is in line with normative data provided by other Italian studies (Santangelo et al., 2015; Aiello et al., 2022). Besides, due to the limited availability of validated socio-cognitive tests, other sub-components of social cognition were not assessed. Finally, no motor scores including hypomimia were included in this study, thus preventing to evaluate the role of reduced facial mimicry in social tasks, particularly for emotion recognition (Künecke et al., 2014; Prenger and Macdonald, 2018).

In conclusion, the results of the present study support that alterations in affective recognition and attribution may occur in PD from the earliest stages of the disease. In agreement with previous evidence (Czernecki et al., 2021), we highlighted the presence of a PD subgroup with socio-cognitive dysfunctions, which in a small percentage of patients represented an isolated deficit, overall supporting the importance of including social cognition in PD neuropsychological assessment. The relevance of socio-cognitive evaluation in clinical practice has been recently underlined in different neurological populations (Henry et al., 2016; Cotter et al., 2018; Dodich et al., 2020), also in consideration of the significant consequences of these deficits in social integration, well-being and quality of life (Bodden et al., 2010; Martinez et al., 2018; Dodich et al., 2021b). The inclusion of social tasks in the cognitive assessment of PD, as well as in the evaluation of MCI due to other neurodegenerative diseases (e.g., Alzheimer’s Disease) (Boccardi et al., 2021), will promote the full characterization of these deficits, as well as their clinical role in a diagnostic and prognostic framework.

The raw data supporting the conclusions of this article will be made available by the authors, without undue reservation.

The studies involving human participants were reviewed and approved by University of Trento Ethical Committee. The patients/participants provided their written informed consent to participate in this study.

AD, CP, LT, and FZ: conception and organization of the project. MP, CL, MM, and RD: data acquisition. AD, CM, and GF: data interpretation and data analyses. AD and GF: first drafting of the work. All authors revised the manuscript and provided the approval of the work.

The authors declare that the research was conducted in the absence of any commercial or financial relationships that could be construed as a potential conflict of interest.

All claims expressed in this article are solely those of the authors and do not necessarily represent those of their affiliated organizations, or those of the publisher, the editors and the reviewers. Any product that may be evaluated in this article, or claim that may be made by its manufacturer, is not guaranteed or endorsed by the publisher.