Social Cognition in Patients with Amnestic Mild Cognitive Impairment and Mild Dementia of the Alzheimer Type

Abstract

Background:

Social cognition (SC) is a core criterion for neurocognitive disorders. However, findings in patients with amnestic mild cognitive impairment (aMCI) and dementia of the Alzheimer type (DAT) are inconsistent.

Objective:

We report assessments of emotion recognition (ER), affective and cognitive theory of mind (ToM) in young (YC) and older controls (OC) compared to aMCI and DAT.

Methods:

28 aMCI, 30 DAT, 30 YC, and 29 OC received tests of SC and a comprehensive neuropsychological assessment. Analysis of covariance was used to determine group differences. Multiple regression models were applied to identify predictors for each SC task.

Results:

In controls, OC performed worse in ER and both ToM tasks compared to YC except for one subtest. No significant differences were found between OC and patients concerning ER and affective ToM. In cognitive ToM, differences between OC and patients depended on content and cognitive load with significant impairment in DAT compared to OC. A cognitive composite score predicted SC in OC, but not in patients. Associations of SC with single cognitive domains were found in all groups with language and complex attention as best predictors. Not all variance of SC performance was explained by variance in cognitive domains.

Conclusion:

Lower performance on SC tasks in OC versus YC was confirmed, although not all tasks were equally affected. With progressive cognitive impairment, cognitive ToM is more impaired than ER or affective ToM. SC seems to be at least partly independent of other cognitive domains, justifying its inclusion in batteries for dementia diagnostic.

Keywords

Amnestic mild cognitive impairment dementia of the Alzheimer type emotion recognition social cognition theory of mind

INTRODUCTION

Social cognition (SC) has attracted increasing attention in diagnostics and research in neurocognitive disorders, such as Alzheimer’s disease [1]. This is partly driven by the fifth version of the Diagnostic and Statistical Manual of Mental Disorders (DSM-5), which has included SC as core criterion for neurocognitive disorders besides established domains, i.e., learning and memory, language, complex attention, executive functions, and visuo-constructive and visuo-perceptual abilities [2].

According to DSM-5, SC can be divided into the components emotion recognition (ER) and theory of mind (ToM). Regarding the first, basic emotions are automatically and cross-culturally recognized and probably rely on an innate mechanism [3] which is distinct from more complex emotions [4]. Facial ER can be conceptualized as a ToM precursor [5], as a basic aspect of affective ToM [6] or as an underlying process for empathy [7].

ToM itself can be divided into two subcomponents: cognitive and affective ToM. Cognitive ToM describes the understanding of cognitive states, beliefs, thoughts, or intentions of other people [6, 8] and can be assessed by different tests such as tests of attribution of intention or of false belief. Within cognitive ToM, first order, second order, and third order mental representations can be differentiated. Affective ToM on the other hand comprises the understanding of affective states, emotions or feelings of others [8]. It is often addressed with the “Reading the Mind in the Eyes” Test - Revised Version (RMET) [9]. Here, the understanding of basic and also more complex emotions that can be considered as blends of mental states, are assessed [4].

SC influences the ability to act and communicate appropriately with other individuals in social situations of variable complexity. There is a consensus that SC declines with increasing age in both, ER as well as ToM [4 , 10–14]. Many authors argue that more basic cognitive functions that contribute to SC, such as executive functions, may be responsible for this age-related loss [4, 15]. In the last years, there is a growing body of literature investigating SC in mild cognitive impairment (MCI) and dementia of Alzheimer type (DAT) with inconclusive results. Heterogeneous definitions of SC (also of relevant subcomponents, including empathy [16]), different tests of SC, non-standardized test procedures, interpretation of raw scores without normative data, small sample sizes and sample differences (e.g., stage of disease, age at onset, different cognitive profile) make it difficult to compare and generalize findings. Only recently, systematic reviews and meta-analyses have tried to create a more unambiguous picture. Nevertheless, comparisons across studies are biased by relevant confounders like age, gender distribution, years of education or cognitive profiles [17]. With regard to ER, two systematic reviews on facial expression recognition in DAT show divergent results ranging from no group differences to severe differences for recognition of specific emotions [18, 19]. Performance in DAT seems to depend on the tasks’ basic cognitive requirements, the presented emotional intensity and the severity of the disease [19, 20]. With regard to ToM, a recent meta-analysis [17] demonstrated significant impairment, particularly in advanced tasks in DAT. While most research did not compare performance of affective and cognitive ToM within one study, two recent meta-analysis demonstrated that relative to controls, DAT is associated with marked deficits in affective as well as in cognitive ToM [16, 17].

Research has additionally focused on MCI. MCI patients show cognitive impairment without significant functional decline. People diagnosed with amnestic MCI (aMCI) have an impaired memory function and an elevated risk for conversion to DAT [21]. Results on SC are heterogeneous. Regarding ER, results range from no to severe impairment in MCI/aMCI compared with healthy controls [18 , 22]. Two meta-analyses reported worse performance of ToM ability in MCI/aMCI compared to healthy controls [17, 23]. One of them compared individual ToM task differences and directly contrasted affective and cognitive ToM [17]. Significant impairment was shown for both aMCI and DAT compared to controls with more severe deficits in DAT in most individual tests [17].

The differentiation between ER as well as between affective and cognitive ToM as distinct factors of SC has been shown to be particularly useful in the characterization of patterns of impairment in neurodegenerative diseases like DAT, frontotemporal dementia (FTD), amyotrophic lateral sclerosis, Parkinson’s and Huntington’s diseases [24 –26]. An example is the impaired performance of both affective and cognitive ToM in FTD compared to DAT patients who showed impairment also in cognitive ToM, but much less in affective ToM tasks [27]. Another important aspect that highlights the need to gain more knowledge about impaired SC in aMCI and DAT is its relevance as a key predictor of broader outcomes like mental comorbidities, well-being, social integration, and quality of life [1 , 28]. Also, SC deficits in DAT predict caregiver burden and depression [29].

The influence of different more basic cognitive domains on SC in neurodegenerative diseases is not fully resolved. It remains controversial whether deficits in SC in DAT result from a domain-specific ToM impairment [30, 31] that is distinct from other cognitive functions or whether they are secondary to deficits in other cognitive domains like executive functions, memory, language, or visuo-spatial skills [32]. The latter entails the identification of cognitive predictors for SC and has been supported by recent evidence [27 , 33–35].

Here, we comprehensively compare ER as well as affective and cognitive ToM measures with first, second, and third order complexity levels between young (YC) and old (OC) healthy individuals and in comparison to aMCI and mild DAT patients. We also examine the relationship of SC with global cognition and individual cognitive domains.

We hypothesize that performance in SC differs between YC and OC with reduced performance in OC. Additionally, we expect that with increasing cognitive impairment, deficits in ER and ToM occur (OC > aMCI > DAT). Moreover, in aMCI and DAT we anticipate a relationship between cognitive domains and global cognition with SC.

METHODS

Participants

We included 30 YC, 29 OC, 28 aMCI, and 30 mild DAT patients. All patients were recruited at the Center for Memory Disorders at the University of Cologne, Germany. Mild DAT and aMCI patients were diagnosed according to the clinical National Institute on Aging-Alzheimer’s Association (NIA-AA) criteria [36, 37] based on the available clinical information, including medical history, physical examination, neuropsychological testing, blood laboratory evaluations, and magnetic resonance imaging (MRI). In addition, all participants in the MCI group had an amnestic subtype. Cerebrospinal fluid (CSF) biomarkers were available in a subset, but are not used for the present study, where we defined the patient groups according to the criteria named above.

Control subjects were recruited via advertisement. All neuropsychological assessments of this study, including SC tasks were performed through home visits within three weeks to 12 months after the initial clinical diagnosis was made. All patients maintained their initial syndromal diagnosis. Inclusion criteria for all subjects were ≧ six education years and fluent German language. OC, aMCI, and DAT had to be older than 50 years of age while YC were between 20–35 years old. Subjects were excluded when they (1) had a history of other severe mental disorders or neurological diseases. In addition, the Beck Depression Inventory II (BDI) [38] was used to assess symptoms of depression in YC and the Geriatric Depression Scale (GDS) [39] in OC, aMCI, and mild DAT. Cut-off scores for exclusion were ≧ 20 for the BDI and ≧ 10 for the GDS. Additional exclusion criteria were (2) uncorrected vision or hearing impairment interfering with the testing procedures (in patients, this was recorded in the medical history and documented after neuropsychological assessment; in patients and controls this was assessed at the beginning of the study), and (3) medication, that might interfere with cognition (e.g., benzodiazepines). Control subjects had to be cognitively unimpaired as defined by a score of 28 points or higher on the Mini-Mental-State-Examination (MMSE) [40].

All participants provided written informed consent prior to study participation. The study was approved by the institutional review board (Reference nr: 17–115).

Assessments

Demographics, medical history, and general cognition

Demographic information and medical history were obtained by questionnaires. The overall cognitive state was determined with the MMSE.

Neuropsychological assessments

All participants completed a neuropsychological test battery. The test selection was guided by the cognitive domains defined through DSM-5, which are learning and memory, complex attention, executive functions, language, as well as perceptual-motor functions. Learning and memory was assessed using the Logical Memory test with immediate and delayed recall (Wechsler Memory Scale WMS-IV version 1 for participants age < 65 years, version 2 for participants > 65 years) [41]. Complex attention was assessed with the Trail-Making-Test (TMT)-A and -B [42]. Executive function was assessed with the TMT B/A [43]. Language was assessed with verbal fluency measures (phonematic (“s-words”) and semantic (animals) one-minute word fluency) [44]. Finally, perceptual-motor function was assessed with the Luria sequence [45].

Emotion recognition

To study facial emotion recognition, stimuli of the sets of the Karolinska Directed Emotional Faces (KDEF) were presented to the participants [46]. These stimuli are colored photographs of faces from different males and females who display six basic emotions as well as a neutral facial expression. The basic emotions include happiness, sadness, fear, disgust, anger, and surprise. For every emotion, a female and a male stimulus with higher (100%) and lower (75%) probability of recognition were chosen based on normative data from the validated set [47]. Thus, the participants were presented four stimuli per emotion. The final test included 28 colored photographs displayed in a pseudo-random sequence (same sequence for every participant). The stimuli were provided as paper prints. All basic emotions were written as words next to the picture of the face and the participants were asked to select the best matching option. Participants were instructed that the same emotion could occur repeatedly. “Surprise” was defined as negative fright instead of positive surprise. The maximum score of this task was 28, which was achieved when all facial emotions were correctly identified.

Affective ToM

To assess affective ToM, the “Reading the Mind in the Eyes Test Revised Version” (RMET) [48] based on the work of Baron-Cohen et al. was administered [9]. Subjects are presented 36 different pairs of eyes with four mental state labels written next to them on paper cards. Participants are asked to select which emotion label corresponds to the emotion that the eyes are expressing. The set included basic as well as more complex emotions. The maximum score was 36. Higher scores reflect better performance.

Cognitive ToM

Participants completed the “Theory of Mind Picture Stories Task” created by Brüne [49]. The test contains six picture stories, each consisting of four colored pictures in cartoon style. Each story is about an interaction of several people cooperating or deceiving each other in order to reach a goal. The task to perform is rearranging the initially randomly mixed pictures in a meaningful order. If the cards are not arranged correctly, the investigator provides this information to the participant and corrects the order in front of the subject. In our setup, we did not measure the time of completion in order to avoid confounding by reduction in motor speed in the older participants. Once the cards were brought in a logical order, the participants were asked to answer two to five open questions about the pictured story. Besides two simple control questions for verifying the comprehension of the story (reality check), these questions covered different levels of ToM complexity (first, second and third order) as well as recognition of cheating, reciprocity, and deception. Two measures of cognitive ToM can be obtained, which are picture sorting (max. 36 points) and response to questions (max. 23 points). Both are summed up to a maximum total score of 59 points. Higher scores reflect better performance.

Data analysis

Statistical analysis was performed using IBM SPSS Statistics, Version 22 (IBM Corp., 2013). A threshold of p < 0.05 was applied to define statistical significance. Demographic (age, education years) and clinical data (depression, MMSE) as well as neuropsychological tests and cognitive composite score (see below) were compared between the four groups applying one-way analysis of variance (ANOVA) with post-hoc Bonferroni test. A composite score was calculated for each subject using the z-transformed results of immediate and delayed recall of logical memory, of the ratio TMT B/A, of semantic fluency and of the Luria sequence. These tests were chosen according to the literature on cognitive impairment in DAT [50].

One-way analyses of covariance (ANCOVAs) with post-hoc Bonferroni tests were used to analyze group differences in ER as well as in affective and cognitive ToM. Years of education served as a covariate. For ER, we analyzed the total number of correctly labelled facial expressions and the number of correctly identified emotional expressions for each basic emotion. For the affective ToM task, the total number of correct answers was analyzed. For the cognitive ToM task, a total score was generated which comprised the scores of the two subtests, i.e., card sorting and answers to questions. Both subtests were also analyzed separately. The subtest answering questions was further divided into first, second, third order questions and questions of reality, reciprocity, deception, and cheating detection.

Next, we analyzed whether the cognitive composite score predicted ER, affective or cognitive ToM performance applying linear regression models for the three older groups together and for each of the groups separately. YC were excluded from this analysis, because they performed significantly different compared to OC in most of the ToM tasks.

Finally, stepwise multiple regression models were applied to test which of the eight individual cognitive test scores (TMT-A, TMT-B, TMT B/A, semantic and phonematic fluency, Luria sequence, logical memory immediate and delayed recall) best predicted ER, affective or cognitive ToM performance in the older groups (OC, aMCI, DAT).

RESULTS

Demographic and neuropsychological data

Results for demographic and domain-specific neuropsychological tests are listed in Table 1. The groups did not differ with regard to gender distribution. OC and both patient groups did not differ with regard to age. The YC group had significantly more education years than all other groups. For this reason, education years served as a covariate in all ANCOVAs which included YC. Depression questionnaires showed significantly higher scores in DAT and aMCI compared to OC although all scores were below the cut-off for clinically relevant depression. YC and OC had significantly higher MMSE scores than aMCI and DAT patients. Additionally, aMCI patients showed significantly higher MMSE scores than DAT patients.

Table 1

Demographic information and cognitive data for all groups

		YC N = 30	OC N = 29	aMCI N = 28	DATN = 30	χ²	p	Direction^a
Demographics		N (%)	N (%)	N (%)	N (%)
	Males	15 (50.00)	9 (31.03)	14 (50.00)	16 (53.33)	3.64	0.304
		Mean (SD)	Mean (SD)	Mean (SD)	Mean (SD)	F
	Age, y	24.27 (2.60)	68.38 (11.84)	73.64 (5.76)	72.07 (8.66)	261.17	< 0.001	YC < OC, aMCI, DAT
	Education, y	16.47 (1.55)	13.90 (4.23)	12.39 (2.94)	13.73 (3.47)	8.32	< 0.001	YC > OC, aMCI, DAT
	MMSE	29.77 (0.50)	29.31 (0.85)	26.64 (1.99)	25.33 (2.70)	43.94	< 0.001	YC > aMCI, DAT; OC > aMCI, DAT; aMCI > DAT
	GDS/BDI	BDI 3.77 (3.61)	GDS 1.10 (1.01)	GDS 2.46 (2.35)	GDS 2.40 (1.94)	4.998	0.009	OC < aMCI, DAT
DSM-5 domains
Learning and memory	immediate recall LM^*Δ	9.50 (3.08)	8.69 (3.26)	6.32 (3.72)	4.17 (2.77)	13.96	<0.001	DAT < YC, OC, aMCI
	delayed recall LM^*Δ	9.07 (3.33)	7.90 (2.97)	4.79 (4.26)	1.60 (1.30)	30.45	< 0.001	DAT < YC, OC, aMCI; aMCI < YC, OC
Complex attention	TMT-A (s)	20.57 (6.60)	49.62 (31.22)	66.64 (30.24)	70.27 (40.08)	11.21	< 0.001	YC < OC, aMCI, DAT; OC < DAT
	TMT-B (s)⁺	47.07 (14.48)	108.90 (67.85)	177.64 (84.80)	214.87 (83.87)	27.31	< 0.001	YC < OC, aMCI, DAT; OC < aMCI, DAT
Executive function	TMT B/A^Δ	2.38 (0.68)	2.27 (0.71)	2.94 (1.64)	3.60 (1.83)	5.71	< 0.001	DAT > YC, OC
Language	Phonematic fluency	17.27 (4.97)	13.00 (4.37)	10.94 (4.39)	10.40 (4.27)	9.01	< 0.001	YC > OC, aMCI, DAT
	Semantic fluency^Δ	28.43 (4.92)	23.70 (8.07)	17.07 (6.32)	13.67 (3.78)	28.90	< 0.001	YC > aMCI, DAT; OC > aMCI, DAT
Perceptual-motor	Luria Sequence function^Δ	2.93 (0.25)	2.69 (0.54)	2.25 (0.70)	1.93 (0.94)	11.65	< 0.001	YC > aMCI, DAT; OC > DAT
Cognitive composite score		0.72 (0.36)	0.34 (0.51)	–0.27 (0.67)	–0.81 (0.39)	55.39	< 0.001	YC > OC, aMCI, DAT; OC > aMCI, DAT; aMCI > DAT

YC, young controls; OC, older controls; aMCI, amnestic mild cognitive impairment; DAT, Dementia of the Alzheimer type; MMSE, Mini-Mental State Examination; GDS, Geriatric Depression Scale; BDI, Beck Depression Inventory; LM, Logical memory; TMT, trail making test. ^* LM, Logical Memory WMS-IV scaled score points. ⁺ TMT-B, max. duration 300 s. Composite score, mean of z-values for neuropsychological variables marked with ^Δ. ^aSignificant post-hoc result (Bonferroni).

As expected, neuropsychological assessments showed lowest scores in the DAT group and best performance in YC. The cognitive composite score discriminated between all groups.

Social cognition

Emotion recognition performance across groups

The ANCOVA revealed a significant group effect for the total score of facial ER (F(3,112) = 6.729, p < 0.001) (see Table 2). Post-hoc tests showed that YC were significantly better than all other groups, but there were no differences between OC, aMCI, and mild DAT patients (see Fig. 1A). Concerning the different basic emotions, significant group differences were found for anger (F(3,112) = 9.33, p < 0.001) with significant post-hoc effects between YC and both patient groups (YC > aMCI, DAT) and for sadness (F(3,112) = 8.336, p < 0.001) with significant post-hoc differences for YC in comparison to all other groups (YC > OC, aMCI, DAT). The analysis revealed no significant group effects for faces showing happy, surprised, fearful, disgusted, and neutral expression.

Table 2

Results for social cognition (emotion recognition, affective and cognitive ToM) and comparison between groups

Social cognition		YC Mean (SD)	OC Mean (SD)	aMCI Mean (SD)	DAT Mean (SD)	F	p	Direction^a
Emotion recognition	Total correct (max. 28)	25.03 (1.77)	21.66 (3.73)	20.96 (3.19)	20.43 (4.28)	6.73	< 0.001	YC > OC, aMCI, DAT
	Happiness (max. 4)	3.87 (0.35)	3.83 (0.38)	3.86 (0.36)	3.57 (0.73)	2.81	0.043
	Surprise (max. 4)	3.93 (0.25)	3.45 (0.95)	3.54 (0.84)	3.50 (0.78)	1.23	0.301
	Anger (max. 4)	3.80 (0.41)	3.07 (1.07)	2.36 (1.10)	2.43 (1.28)	9.33	< 0.001	YC > aMCI, DAT
	Fear (max. 4)	2.77 (1.19)	2.07 (1.19)	2.11 (0.96)	2.40 (1.33)	1.23	0.302
	Sadness (max. 4)	3.70 (0.47)	2.69 (0.97)	2.36 (1.28)	2.43 (1.01)	8.34	< 0.001	YC > OC, aMCI, DAT
	Disgust (max. 4)	3.27 (1.02)	3.10 (1.26)	3.18 (1.12)	2.57 (1.25)	2.00	0.118
	Neutral (max. 4)	3.70 (0.79)	3.45 (0.87)	3.57 (0.74)	3.53 (1.01)	0.28	0.837
Affective ToM	Total correct (max. 36)	26.77 (2.91)	19.86 (5.40)	18.79 (3.65)	17.23 (4.68)	20.68	< 0.001	YC > OC, aMCI, DAT
Cognitive ToM	Sum ToM Sorting and Questions ^# (max. 56)	55.07 (1.84)	45.62 (8.39)	41.14 (8.44)	37.17 (11.40)	18.20	< 0.001	YC > OC, aMCI, DAT; OC > DAT
	Sorting (max. 36)	35.07 (1.84)	26.55 (7.43)	23.68 (6.81)	20.67 (8.69)	17.74	< 0.001	YC > OC, aMCI, DAT; OC > DAT
	Questions (without 3^rd order) (max. 20)	20.00 (0.00)	19.07 (1.51)	17.46 (2.62)	16.50 (3.67)	9.30	< 0.001	DAT < YC, OC
	1^st order (max 5)	5 (0.00)	4.76 0.51)	4.68 (0.67)	4.53 (1.07)	1.28	0.286
	2^nd order (max. 5)	5 (0.00)	4.69 (0.66)	3.96 (1.23)	3.77 (1.36)	6.97	< 0.001	DAT < YC, OC
	3^rd order (max. 3)	3 (0.00)	2.66 (0.72)	n.a.	n.a.	2.07	0.156
	Reality (max. 2)	2 (0.00)	1.97 (0.19)	1.93 (0.26)	1.83 (0.38)	2.11	0.103
	Reciprocity (max. 3)	3 (0.00)	2.86 (0.35)	2.79 (0.50)	2.73 (0.58)	1.66	0.181
	Deception (max. 3)	3 (0.00)	2.79 (0.56)	2.46 (0.69)	2.30 (0.95)	4.60	0.004	DAT < YC, OC
	Cheating detection (max. 2)	2 (0.00)	2.00 (0.00)	1.71 (0.54)	1.33 (0.80)	11.99	< 0.001	DAT < YC, OC, aMCI

YC, young controls; OC, older controls; aMCI, amnestic mild cognitive impairment; DAT, dementia of the Alzheimer type; ToM, theory of mind. ^#Scoring without 3rd order items. ^aSignificant post-hoc result (Bonferroni).

Fig. 1

A) Results for emotion recognition (max. 28). Young controls (YC) were significantly better than older controls (OC), patients with amnestic mild cognitive impairment (aMCI), and patients with mild dementia of the Alzheimer type (DAT). B) Results for the affective Theory of mind (ToM) task (max. 36). YC were significantly better than OC, patients with aMCI, and patients with mild DAT. C) Results for the cognitive ToM task (sum of card sorting and questions, max. 56). YC were significantly better than OC, patients with aMCI, and patients with mild DAT. DAT patients performed significantly worse than YC and OC.

Affective ToM performance across groups

Concerning affective ToM, the ANCOVA showed a significant group effect for the total score (F(3,112) = 20.680, p < 0.001) (see Table 2). Significant difference between YC and all other groups were revealed with post-hoc Bonferroni tests (see Fig. 1B).

Cognitive ToM performance across groups

During data acquisition, it became obvious that answering third order ToM questions was too difficult for both patient groups and only very few participants provided answers. Therefore, these results were not further analyzed in the aMCI and DAT group. This resulted in a maximum score of 20 instead of 23 points for the subtest response to questions and a maximum total score for cognitive ToM in these groups of 56 instead of 59 points.

The ANCOVA revealed a significant group effect for the total score (F(3,112) = 18.200, p < 0.001) (see Table 2). Post-hoc Bonferroni tests showed a significant difference between YC and all other groups. Furthermore, significant differences were found between OC and DAT patients (see Fig. 1C). The subtest of card sorting also showed a significant group effect (F(3,112) = 17.736, p < 0.001). Post-hoc tests showed significant differences between YC and all three other groups as well as between OC and DAT. The subtest on answering questions (without third order questions) showed a significant group effect (F(3,112) = 9.297, p < 0.001). Significant post-hoc effects revealed worse performance in DAT patients compared to YC and OC.

In separate analyses with regard to the level of cognitive load, ANCOVAs did not show significant differences between the four groups for first order questions, but for second order questions (F(3,112) = 6.972, p < 0.001). Again, post-hoc analyses showed that DAT patients were more impaired than YC and OC. There were no significant differences between aMCI and mild DAT.

All subjects were able to answer at least one of the two control questions (“reality check”) of the cartoon picture stories correctly supporting integrity of cognitive abilities on a basic level. ANCOVAs did not show significant differences between the four groups for reality or reciprocity questions. Questions assessing deception showed significant differences in the ANCOVA (F(3,112) = 4.599, p = 0.004). Significant post-hoc effects revealed worse performance in DAT patients compared to YC and OC. The ANCOVA additionally showed significant differences between the four groups for cheating detection (F(3,112) = 11.989, p < 0.001). Post-hoc analyses indicated that DAT patients were more impaired than YC, OC, and aMCI.

Linear regression models for emotion recognition by cognitive composite score

The linear regression models for ER are listed in Table 3. The adjusted R ² for the overall model (including OC, aMCI, and DAT) was 0.11, indicative for a moderate goodness-of-fit, i.e., 11% of variance can be explained by the cognitive composite score. The regression results on single group level were only significant for OC, where the composite score predicted ER performance by 46% (p < 0.001). For aMCI and DAT, no significant results were found.

Table 3

Linear regression models for emotion recognition, affective and cognitive ToM with the cognitive composite score as predictor

		R²	Adjusted R²	β	B	95% CI	p
Emotion recognition	Overall model	0.12	0.11	0.350	1.861	0.79–2.94	< 0.001
	OC	0.48	0.46	0.690	5.061	2.96–7.16	< 0.001
	aMCI	0.02	0.00	0.154	0.734	–1.17–2.63	0.453
	DAT	0.08	0.05	0.286	3.118	–0.92–7.16	0.125
Affective ToM	Overall model	0.12	0.11	0.344	2.293	0.94–3.64	0.001
	OC	0.17	0.14	0.411	4.358	0.54–8.18	0.027
	aMCI	0.02	–0.02	0.138	0.750	–1.43–2.93	0.485
	DAT	0.07	0.04	0.261	3.106	–1.35–7.56	0.164
Cognitive ToM Sorting and Questions	Overall model	0.27	0.26	0.515	7.322	4.69–9.95	< 0.001
	OC	0.44	0.42	0.663	10.935	6.07–15.81	< 0.001
	aMCI	0.14	0.11	0.374	4.713	0.00–9.43	0.050
	DAT	0.08	0.05	0.282	8.185	–2.59–18.96	0.131
Cognitive ToM Sorting	Overall model	0.23	0.23	0.484	5.466	3.33–7.60	< 0.001
	OC	0.42	0.40	0.649	9.480	5.09–13.87	< 0.001
	aMCI	0.13	0.10	0.364	3.709	–1.12–7.53	0.057
	DAT	0.05	0.02	0.230	5.089	–3.24–13.41	0.221
Cognitive ToM Questions	Overall model	0.20	0.19	0.448	1.856	1.06–2.66	< 0.001
	OC	0.24	0.21	0.490	1.455	0.43–2.48	0.007
	aMCI	0.07	0.03	0.257	1.005	–0.52–2.53	0.187
	DAT	0.11	0.08	0.331	3.096	–0.32–6.51	0.074

YC, young controls; OC, old controls; aMCI, amnestic mild cognitive impairment; DAT, dementia of the Alzheimer type; ToM, theory of mind.

Linear regression models for affective ToM by cognitive composite score

For the affective ToM total score, the cognitive composite score predicted 11% of variance, indicative of a moderate goodness-of-fit (see Table 3). The regression results on single group level were only significant for OC with an adjusted R ² of 0.14 (p = 0.027). There were no significant results for aMCI and DAT.

Linear regression models for cognitive ToM by cognitive composite score

Concerning cognitive ToM (total score, including “sorting” and “questions”), the cognitive composite score predicted 26% of variance, indicative for a high goodness-of-fit (adjusted R² = 0.26). Also, the prediction of ToM performance by the cognitive composite score was significant in OC (42%, p < 0.001) but not in aMCI and DAT patients (see Table 3). When “sorting” and “questions” were analyzed separately, the composite score significantly explained 23% of “sorting” and 19% of “questions” when the three groups were included. The subtest “sorting” was also predicted by the composite score in OC (40%, p < 0.001) but not in aMCI and DAT patients. Similarly, the subtest “questions” was only significantly associated with the composite score in OC (21%, p = 0.007).

Stepwise multiple regression model for emotion recognition

Results for the stepwise multiple regression models are displayed in Table 4. When the three groups were considered together, phonemic word fluency and complex attention (TMT-A) were identified as significant predictors and explained 28% of variance for ER. While semantic word fluency (adjusted R² = 0.50) was found as significant predictor in OC, worse TMT-A results (adjusted R² = 0.17) were significantly associated with impaired ER in aMCI patients. In DAT patients, again, phonemic word fluency served (adjusted R² = 0.40) as significant predictor.

Table 4

Stepwise multiple regression models for emotion recognition, affective and cognitive ToM with individual test scores

		Significant variables	Δ R²	Δ	Total adjusted R²	β	B	95% CI	p
Emotion recognition	Overall model	1. Phonemic word fluency	0.24	0.23	0.23	0.488	0.414	0.26–0.57	< 0.001
		2. TMT-A (s)	0.06	0.05	0.28	–0.258	–0.028	–0.05––0.01	0.010
	OC	1. Semantic word fluency	0.52	0.50	0.50	0.719	0.333	0.21–0.46	< 0.001
	aMCI	1. TMT-A (s)	0.20	0.17	0.17	–0.442	–0.047	–0.09––0.01	0.018
	DAT	1. Phonemic word fluency	0.42	0.40	0.40	0.650	0.652	0.36–0.95	< 0.001
Affective ToM	Overall model	1. Semantic word fluency	0.19	0.18	0.18	0.435	2.258	1.25–3.27	< 0.001
	OC	1. Semantic word fluency	0.26	0.23	0.23	0.506	2.799	0.91–4.68	0.005
	aMCI	No significant predictors
	DAT	1. Semantic word fluency	0.17	0.14	0.14	0.406	4.153	0.533–7.77	0.026
Cognitive ToM	Overall model	1. Semantic word fluency	0.33	0.32	0.32	0.574	0.769	0.53–1.00	< 0.001
Sorting and Questions		2. TMT-A (s)	0.09	0.09	0.41	–0.329	–0.94	–0.15––0.04	< 0.001
	OC	1. Phonemic word fluency	0.54	0.52	0.52	0.733	1.408	0.89–1.92	< 0.001
		2. Semantic word fluency	0.09	0.08	0.60	0.421	0.438	0.09–0.79	0.016
	aMCI	1. TMT-A (s)	0.18	0.14	0.14	–0.418	–0.117	–0.22––0.01	0.027
	DAT	1. Phonemic word fluency	0.35	0.33	0.33	0.590	1.574	0.74–2.41	0.001

YC, young controls; OC, old controls; aMCI, amnestic mild cognitive impairment; DAT, dementia of the Alzheimer type; ToM, theory of mind.

Stepwise multiple regression model for affective ToM

The model containing all three groups revealed semantic word fluency with 18% as a significant predictor for affective ToM (see Table 4). Semantic fluency was also a significant predictor in OC and DAT explaining 23% and 14% variance. There was no significant predictor for aMCI.

Stepwise multiple regression models for cognitive ToM

When the whole group of older participants was analyzed, semantic word fluency and complex attention (TMT-A) predicted cognitive ToM performance significantly by 41%. Group-wise regressions identified phonemic word fluency and TMT-A as significant predictors in OC (total adjusted R² = 0.60). While the performance on the TMT-A was significantly associated with impaired cognitive ToM in aMCI patients (adjusted R² = 0.14), again, phonemic word fluency predicted cognitive ToM in DAT patients best (adjusted R² = 0.33, see Table 4).

DISCUSSION

In this study, we measured ER as well as affective and cognitive ToM performance in healthy YC and OC and in aMCI and mild DAT patients. Results showed distinct patterns of impairment in each group with YC showing best performance, followed by OC, aMCI, and DAT. Cognitive ToM measures found more pronounced differences between groups than ER and affective ToM tasks. The cognitive composite score significantly predicted SC only in OC, but not in patients. Cognitive domains of language and complex attention predicted ER and ToM performance in OC and both patient groups except for affective ToM in aMCI.

Emotion recognition

ER is described as an important ToM prerequisite [5]. Older adults show reduced performance in recognizing basic emotions across different modalities [14]. This has been attributed to age-related structural and functional neuronal changes [51]. Recognition of sad, angry, and fearful faces seem to be reduced compared to preserved recognition of disgust and happiness [51, 52]. Our data demonstrate significantly better global ER performance in YC than in OC and recognition of sadness in particular. In general, we observed high recognition rates. Including pictures with more variance in emotional intensity might facilitate the identification of differences between groups [18, 19].

When looking at the performance of the two patient groups compared to OC, there was no significant difference in global ER. On the numerical level and in accordance with a recent systematic review [18], recognition of happy faces in DAT was better than that of other basic emotions, whereas recognition of sadness and anger was most difficult. Additionally, recognition of fear was similarly reduced in our data. Visual scanning of important information from different facial areas is required to identify and label the expressed emotion, which involves attention, social knowledge, and language. In aMCI and DAT, ER was best predicted by TMT-A and phonemic word fluency. These tasks cannot be fully assigned to one single cognitive domain. Instead they are classified as measures of language skills, attention processes and executive functions, all known to influence SC [53, 54]. These results suggest that ER is in part related to these specific cognitive capacities [18].

Affective ToM

The results for the affective ToM task were comparable to the ER results, which supports data showing common neuronal mechanism of both [55]. There was a significant group effect showing better performance in YC than in OC, aMCI, and DAT with no differences between the latter groups. As such, our results are in line with single studies [56, 57] but do not replicate the results of a meta-analysis [17]. It has been shown before that the performance in this task declines with age [58] which might offer an explanation for the inconsistent findings between studies.

Semantic fluency explained 23% of variance in this task in the OC and 14% in the DAT group, which is in accordance with a recent work showing association of affective ToM with verbal abilities [33, 58]. The association did not reach significance in the aMCI group. One limitation of the RMET, which is one of the most frequently used tests for assessing affective ToM in adults, is, that it can be considered more of an index of emotion recognition rather than of ToM ability [59].

Cognitive ToM

As expected, cognitive ToM (total score, subtest card sorting) was reduced in OC compared to YC. However, our results demonstrated that this reduction does not affect all aspects of cognitive ToM since no differences between YC and OC were found for the task of answering questions. YC reached ceiling effects in all questions, which may contribute to the lack of a significant difference, while OC performed only mildly worse. Answering questions about thoughts, beliefs and intentions of others seems to require abilities that tend to be stable over the life span like social knowledge and autobiographical memories of social situations [60] as well as language expression. However, age effects may contribute to lower cognitive ToM performance [12]. We observed an association of cognitive ToM with cognitive performance as assessed with a cognitive composite score in OC. Semantic and phonemic word fluency were the best predictors. The amount of explained variance in cognitive ToM for OC was higher for card sorting (40%) compared with answering questions (21%).

Concerning patient groups, DAT patients were significantly more impaired in cognitive ToM than OC (total score, subtasks of sorting cards and answering questions). The aMCI group reached intermediate scores between OC and DAT in all subscores of cognitive ToM which reached significance for cheating detection only. As for ER and affective ToM, no significant associations were found between the cognitive composite score and cognitive ToM in aMCI and DAT. This result is surprising since most studies postulate that ToM impairments reflect a general decline in cognition [32]. Lucena et al. [61] proposed that ToM abilities in DAT corresponds to the profile of cognitive impairment of the individual patient rather than being a general feature of the disease.

The stepwise regression analysis showed a significant relationship between semantic and phonemic word fluency and cognitive ToM in OC. In affective ToM, but even more so in cognitive ToM the amount of explained variance was high (23% and 60%), confirming previous studies describing an impact of basic cognitive domains on SC in healthy subjects [62, 63] with the most robust association being between SC and executive functions [4]. Accordingly, a recent study showed that executive functioning skills predicted SC performance in different tasks of ER, affective and cognitive ToM in an overall model containing older controls and DAT patients [64]. However, when analyzed separately, the association was only significant for controls, not for DAT patients. Further research is needed to better understand the association between different ToM tasks and specific cognitive/executive domains. Overall, our data demonstrate that cognitive domains influence SC in controls and patients. However, basic cognitive functions do not fully explain SC performance in our data [32, 33].

Studies on cognitive ToM in DAT involved different levels of cognitive load. Results were inconsistent, but suggested an association of performance with advanced stages of the disease [65]. Fliss et al. [66] showed that patients with mild DAT performed similarly to healthy controls in first order false belief tasks while having more impaired cognitive performance, whereas patients with moderate DAT showed impaired performance. Laisney et al. [67] supportively reported a correlation between first order false belief questions and the Mattis Dementia Rating Scale measuring general cognitive abilities. In accordance, we found preserved abilities in tasks with low cognitive demands in mild DAT patients, e.g., affective ToM, ER, and first order level of cognitive ToM [27 , 35]. Additionally, we could demonstrate that when cognitive load increased, impairment in SC was detectable as previously described.

Interestingly, these cognitive load effects were not detectable in aMCI patients. Only one significant difference between aMCI and DAT was found in cognitive ToM performance showing better scores for answering questions on cheating detection for aMCI. Other studies reported inconsistent findings ranging from no deficit at all to severe impairment at the stage of aMCI, including ER and affective as well as cognitive ToM tasks [17 , 51]. SC deficits seem to be more severe in multidomain MCI [23], and it is also detectable in non-aMCI patients [22]. The role of these impairments with regard to conversion to dementia needs to be investigated in longitudinal studies [23].

Limitations

An important limitation is the cross-sectional design that prevents conclusions about temporal dynamics and causal relationships. We did not include biomarkers in the present analyses. Future studies may focus specifically on SC in MCI and dementia related to Alzheimer’s disease or other neurodegenerative disorders.

CONCLUSIONS

This is the first study that included YC, OC, aMCI, and DAT to assess ER, affective and cognitive ToM within the same protocol. SC seems to be relatively intact early in the course of DAT compared to other cognitive domains. ER and affective ToM seem to be preserved even longer than cognitive ToM. Increasing cognitive load of cognitive ToM revealed significant differences in performance between controls and patients. Different cognitive domains explained variance of SC partly for all groups. However, SC was associated with overall cognition only in controls, which strengthens the argument for including SC as a standard domain in neuropsychological batteries in DAT.

Footnotes

ACKNOWLEDGMENTS

We thank Prof. Brüne, PD Dr. Bölte and Prof. Litton for the permission to use the ToM task, the RMET and the facial stimuli.

Authors’ disclosures available online ().

References

Henry

, von Hippel

, Molenberghs

, Lee

, Sachdev

(2016) Clinical assessment of social cognitive function in neurological disorders. Nat Rev Neurol 12, 28–39.

American Psychiatric Association (2013) Diagnostic and Statistical Manual of Mental Disorders, American Psychiatric Publishing, Washington, DC.

Izard

(1994) Innate and universal facial expressions: Evidence from developmental and cross-cultural research. Psychol Bull 115, 288–299.

Duval

, Piolino

, Bejanin

, Eustache

, Desgranges

(2011) Age effects on different components of theory of mind. Conscious Cogn 20, 627–642.

Sommer

, Döhnel

, Meinhardt

, Hajak

(2008) Decoding of affective facial expressions in the context of emotional situations. Neuropsychologia 46, 2615–2621.

Coricelli

(2005) Two-levels of mental states attribution: From automaticity to voluntariness. Neuropsychologia 43, 294–300.

Shamay-Tsoory

(2011) The neural bases for empathy. Neurosci 17, 18–24.

Brothers

, Ring

(1992) A neuroethological framework for the representation of minds. J Cogn Neurosci 4, 107–118.

Baron-Cohen

, Wheelwright

, Hill

, Raste

, Plumb

(2001) The “Reading the Mind in the Eyes” Test Revised Version: A study with normal adults, and adults with Asperger syndrome or high-functioning autism. J Child Psychol Psychiatry 42, 241–251.

10.

Kemp

, Després

, Sellal

, Dufour

(2012) Theory of Mind in normal ageing and neurodegenerative pathologies. Ageing Res Rev 11, 199–219.

11.

Moran

(2013) Lifespan development: The effects of typical aging on theory of mind. Behav Brain Res 237, 32–40.

12.

Sandoz

, Démonet

, Fossard

(2014) Theory of mind and cognitive processes in aging and Alzheimer type dementia: A systematic review. Aging Ment Health 18, 815–827.

13.

Grainger

, Rakunathan

, Adams

, Canty

, Henry

(2021) An assessment of age differences in theory of mind using the virtual assessment of mentalizing ability. Aging, Neuropsychol Cogn 28, 97–107.

14.

Henry

, Phillips

, Ruffman

, Bailey

(2013) A meta-analytic review of age differences in theory of mind. Psychol Aging 28, 826–839.

15.

German

, Hehmann

(2006) Representational and executive selection resources in ‘theory of mind’: Evidence from compromised belief-desire reasoning in old age. Cognition 101, 129–152.

16.

Demichelis

, Coundouris

, Grainger

, Henry

(2020) Empathy and Theory of Mind in Alzheimer’s disease: A meta-analysis. J Int Neuropsychol Soc 26, 963–977.

17.

, Zhao

, Zhang

, Shi

, Zhong

, Pan

(2020) Theory of mind in Alzheimer’s disease and amnestic mild cognitive impairment: A meta-analysis. Neurol Sci 41, 1027–1039.

18.

Torres Mendonça De Melo Fádel

, Santos De Carvalho

, Belfort Almeida Dos Santos

, Dourado

MCN

(2019) Facial expression recognition in Alzheimer’s disease: A systematic review. J Clin Exp Neuropsychol 41, 192–203.

19.

Elferink

MW-O

, van Tilborg

, Kessels

RPC

(2015) Perception of emotions in mild cognitive impairment and Alzheimer’s dementia: Does intensity matter? Transl Neurosci 6, 139–149.

20.

Youmans

, Bourgeois

(2010) Theory of mind in individuals with Alzheimer-type dementia. Aphasiology 24, 515–534.

21.

Petersen

, Doody

, Kurz

, Mohs

, Morris

, Rabins

, Ritchie

, Rossor

, Thal

, Winblad

(2001) Current concepts in mild cognitive impairment. Arch Neurol 58, 1985–1992.

22.

Pietschnig

, Aigner-Wöber

, Reischenböck

, Kryspin-Exner

, Moser

, Klug

, Auff

, Dal-Bianco

, Pusswald

, Lehrner

(2016) Facial emotion recognition in patients with subjective cognitive decline and mild cognitive impairment. Int Psychogeriatr 28, 477–485.

23.

Bora

, Yener

(2017) Meta-analysis of social cognition in mild cognitive impairment. J Geriatr Psychiatry Neurol 30, 206–213.

24.

Dodich

, Cerami

, Crespi

, Canessa

, Lettieri

, Iannaccone

, Marcone

, Cappa

, Cacioppo

(2016) Differential impairment of cognitive and affective mentalizing abilities in neurodegenerative dementias: Evidence from behavioral variant of frontotemporal dementia, Alzheimer’s disease, and mild cognitive impairment. J Alzheimers Dis 50, 1011–1022.

25.

Poletti

, Enrici

, Adenzato

(2012) Cognitive and affective Theory of Mind in neurodegenerative diseases: Neuropsychological, neuroanatomical and neurochemical levels. Neurosci Biobehav Rev 36, 2147–2164.

26.

Henry

, Phillips

, von Hippel

(2014) A meta-analytic review of theory of mind difficulties in behavioural-variant frontotemporal dementia. Neuropsychologia 56, 53–62.

27.

Gregory

, Lough

, Stone

, Erzinclioglu

, Martin

, Baron-Cohen

, Hodges

(2002) Theory of mind in patients with frontal variant frontotemporal dementia and Alzheimer’s disease: Theoretical and practical implications. Brain 125, 752–764.

28.

Christidi

, Migliaccio

, Santamaría-García

, Santangelo

, Trojsi

(2018) Social cognition dysfunctions in neurodegenerative diseases: Neuroanatomical correlates and clinical implications. Behav Neurol 2018, 1849794.

29.

Martinez

, Multani

, Anor

, Misquitta

, Tang-Wai

, Keren

, Fox

, Lang

, Marras

, Tartaglia

(2018) Emotion detection deficits and decreased empathy in patients with Alzheimer’s disease and Parkinson’s disease affect caregiver mood and burden. Front Aging Neurosci 10, 120.

30.

Freedman

, Binns

, Black

, Murphy

, Stuss

(2013) Theory of Mind and recognition of facial emotion in dementia. Alzheimer Dis Assoc Disord 27, 56–61.

31.

Leslie

, Thaiss

(1992) Domain specificity in conceptual development: Neuropsychological evidence from autism. Cognition 43, 225–251.

32.

Strikwerda-Brown

, Ramanan

, Irish

(2019) Neurocognitive mechanisms of theory of mind impairment in neurodegeneration: A transdiagnostic approach. Neuropsychiatr Dis Treat 15, 557–573.

33.

Yildirim

, Soncu Buyukiscan

, Demirtas-Tatlidede

, Bilgiç

c B

, Gurvit

(2020) An investigation of affective theory of mind ability and its relation to neuropsychological functions in Alzheimer’s disease. J Neuropsychol 14, 399–415.

34.

Fernandez-Duque

, Baird

, Black

(2009) False-belief understanding in frontotemporal dementia and Alzheimer’s disease. J Clin Exp Neuropsychol 31, 489–497.

35.

Zaitchik

, Koff

, Brownell

, Winner

, Albert

(2006) Inference of beliefs and emotions in patients with Alzheimer’s disease. Neuropsychology 20, 11–20.

36.

Albert

, DeKosky

, Dickson

, Dubois

, Feldman

, Fox

, Gamst

, Holtzman

, Jagust

, Petersen

, Snyder

, Carrillo

, Thies

, Phelps

(2011) The diagnosis of mild cognitive impairment due to Alzheimer’s disease: Recommendations from the National Institute on Aging-Alzheimer’s Association Workgroups on Diagnostic Guidelines for Alzheimer’s Disease. Alzheimers Dement 7, 270–279.

37.

McKhann

, Knopman

, Chertkow

, Hyman

, Jack

, Kawas

, Klunk

, Koroshetz

, Manly

, Mayeux

, Mohs

, Morris

, Rossor

, Scheltens

, Carrillo

, Thies

, Weintraub

, Phelps

(2011) The diagnosis of dementia due to Alzheimer’s disease: Recommendations from the National Institute on Aging-Alzheimer’s Association workgroups on diagnostic guidelines for Alzheimer’s disease. Alzheimers Dement 7, 263–269.

38.

Beck

, Brown

, Steer

(2009) BDI-II Beck' Depressions Inventar Revision, Pearson Assessment and Information GmbH, Frankfurt am Main.

39.

Yesavage

, Sheikh

(1986) Geriatric Depression Scale (GDS): Recent evidence and development of a shorter version. Clin Gerontol 5, 165–173.

40.

Folstein

, Folstein

, McHugh

(1975) ‘Mini-mental State’. A Practical Method for Grading the Cognitive State of Patients for the Clinician. J Psychiatr Res 12, 189–198.

41.

Petermann

, Lepach

(2012) Wechsler Memory Scale - Fourth Edition, German Edition, Pearson Assessment, Frankfurt am Main.

42.

Reitan

(1958) Validity of the Trail Making Test as an indicator of organic brain damage. Percept Mot Skills 8, 271–276.

43.

Arbuthnott

, Frank

(2000) Trail Making Test, Part B as a measure of executive control: Validation using a set-switching paradigm. J Clin Exp Neuropsychol 22, 518–528.

44.

Aschenbrenner

, Tucha

, Lange

(2000) Regens burger Wortflüssigkeits-Test (RWT), Hogrefe Verlag für Psychologie, Göttingen.

45.

Luria

(1980) Higher Cortical Functions in Man, Springer US, Boston, MA.

46.

Lundqist

, Flykt

, Öhman

(1998) The Karolinska Directed Emotional Faces - KDEF, Psychology section, Department of Clinical Neuroscience, Karolinska Institute, Stockholm, Sweden, Stockholm.

47.

Goeleven

, De Raedt

, Leyman

, Verschuere

(2008) The Karolinska Directed Emotional Faces: A validation study. Cogn Emot 22, 1094–1118.

48.

Boelte

Reading Mind in the Eyes Test. Erwachsenenversion. Von Simon Baron-Cohen (2001) Deutsche Bearbeitung von Sven Bölte. http://www.as-tt.de/assets/applets/Augentest_Erwachsene.pdf, Last update 2015, Accessed December 19, 2016.

49.

Brüne

(2003) Social cognition and behaviour in schizophrenia. In The Social Brain: Evolution and Pathology, BrüneM, Ribbert H, SchiefenhövelW, eds. JohnWiley & Sons, Chichester, UK, pp. 277–313.

50.

Donohue

, Sperling

, Salmon

, Rentz

, Raman

, Thomas

, Weiner

, Aisen

(2014) The preclinical Alzheimer cognitive composite: Measuring amyloid-related decline. JAMA Neurol 71, 961–970.

51.

McCade

, Savage

, Naismith

(2011) Review of emotion recognition in mild cognitive impairment. Dement Geriatr Cogn Disord 32, 257–266.

52.

Calder

, Keane

, Manly

, Sprengelmeyer

, Scott

, Nimmo-Smith

, Young

(2003) Facial expression recognition across the adult life span. Neuropsychologia 41, 195–202.

53.

Cho

, Cohen

(2019) Explaining age-related decline in theory of mind: Evidence for intact competence but compromised executive function. PLoS One 14, e0222890.

54.

Shao

, Janse

, Visser

, Meyer

(2014) What do verbal fluency tasks measure? Predictors of verbal fluency performance in older adults. Front Psychol 5, 772.

55.

Mier

, Lis

, Neuthe

, Sauer

, Esslinger

, Gallhofer

, Kirsch

(2010) The involvement of emotion recognition in affective theory of mind. Psychophysiology 47, 1028–1039.

56.

Baglio

, Castelli

, Alberoni

, Blasi

, Griffanti

, Falini

, Nemni

, Marchetti

(2012) Theory of Mind in amnestic mild cognitive impairment: An fMRI study. J Alzheimers Dis 29, 25–37.

57.

Heitz

, Noblet

, Phillipps

, Cretin

, Vogt

, Philippi

, Kemp

, De Petigny

, Bilger

, Demuynck

, Martin-Hunyadi

, Armspach

, Blanc

(2016) Cognitive and affective theory of mind in dementia with Lewy bodies and Alzheimer’s disease. Alzheimers Res Ther 8, 10.

58.

Kynast

, Quinque

, Polyakova

, Luck

, Riedel-Heller

, Baron-Cohen

, Hinz

, Witte

, Sacher

, Villringer

, Schroeter

(2020) Mindreading from the eyes declines with aging - evidence from 1,603 subjects. Front Aging Neurosci 12, 550416.

59.

Oakley

BFM

, Brewer

, Bird

, Catmur

(2016) Theory of mind is not theory of emotion: A cautionary note on the Reading the Mind in the Eyes Test. J Abnorm Psychol 125, 818–823.

60.

Olson

, McCoy

, Klobusicky

, Ross

(2013) Social cognition and the anterior temporal lobes: A review and theoretical framework. Soc Cogn Affect Neurosci 8, 123–133.

61.

Lucena De

, Bhalla

, Belfort Almeida Dos Santos

, Dourado

MCN

(2020) The relationship between theory of mind and cognition in Alzheimer’s disease: A systematic review. J Clin Exp Neuropsychol 42, 223–239.

62.

Charlton

, Barrick

, Markus

, Morris

(2009) Theory of Mind associations with other cognitive functions and brain imaging in normal aging. Psychol Aging 24, 338–348.

63.

Maylor

, Moulson

, Muncer

A-M

, Taylor

(2002) Does performance on theory of mind tasks decline in old age? Br J Psychol 93, 465–485.

64.

Chainay

, Gaubert

(2020) Affective and cognitive theory of mind in Alzheimer’s disease: The role of executive functions. J Clin Exp Neuropsychol 42, 371–386.

65.

Bertoux

, De Souza

, Sarazin

, Funkiewiez

, Dubois

, Hornberger

(2015) How preserved is emotion recognition in Alzheimer disease compared with behavioral variant frontotemporal dementia? Alzheimer Dis Assoc Disord 29, 154–157.

66.

Fliss

, Le Gall

, Etcharry-Bouyx

, Chauviré

, Desgranges

, Allain

(2016) Theory of Mind and social reserve: Alternative hypothesis of progressive Theory of Mind decay during different stages of Alzheimer’s disease. Soc Neurosci 11, 409–423.

67.

Laisney

, Bon

, Guiziou

, Daluzeau

, Eustache

, Desgranges

(2013) Cognitive and affective Theory of Mind in mild to moderate Alzheimer’s disease. J Neuropsychol 7, 107–120.