Impaired Facial Emotion Recognition and Gaze Direction Detection in Mild Alzheimer’s Disease: Results from the PACO Study

Abstract

Background:

Facial emotion recognition (FER) and gaze direction (GD) identification are core components of social cognition, possibly impaired in many psychiatric or neurological conditions. Regarding Alzheimer’s disease (AD), current knowledge is controversial.

Objective:

The aim of this study was to explore FER and GD identification in mild AD compared to healthy controls.

Methods:

180 participants with mild AD drawn from the PACO study and 74 healthy elderly controls were enrolled. Participants were asked to complete three socio-cognitive tasks: face sex identification, recognition of facial emotions (fear, happiness, anger, disgust) expressed at different intensities, and GD discrimination. Multivariate analyses were conducted to compare AD participants and healthy controls.

Results:

Sex recognition was preserved. GD determination for subtle deviations was impaired in AD. Recognition of prototypically expressed facial emotions was preserved while recognition of degraded facial emotions was impacted in AD participants compared to controls. Use of multivariate analysis suggested significant alteration of low-expressed fear and disgust recognition in the AD group.

Conclusion:

Our results showed emotion recognition and GD identification in patients with early-stage AD compared to elderly controls. These impairments could be the object of specific therapeutic interventions such as social cognition remediation or raising awareness of primary caregivers to improve the quality of life of patients with early AD.

Keywords

Alzheimer’s disease elderly facial emotion recognition social cognition

INTRODUCTION

Social cognition (SC) broadly refers to a set of mechanisms allowing individuals of the same species to interact together [1]. Facial emotion processing has been regarded as a central element of SC [2] regarding its crucial aspect as the reflection of a mental state. The misinterpretation of facial emotions could lead to difficulties in social conduct [3] as observed in various major mental illnesses, especially schizophrenia [4] or autism spectrum disorder [5] and in neurodegenerative diseases [6]. Recent advances in neuroimaging have led to a better understanding of the neural networks underlying facial emotion recognition (FER). According to a recent integrative model [7], FER would be based on two separate neural pathways both starting from the early visual occipital cortex. First, the ventral face-processing pathway, centered on the fusiform face area (FFA) in the inferior temporal cortex, which could contribute to the identification of invariable features such as identity, age, and sex. Second, the dorsal face-processing pathway, specialized in the recognition of dynamic elements of the face such as eye-gaze direction or emotions, involving the superior temporal sulcus (STS) and its connections with the amygdala and the orbitofrontal cortex. Considering neurodegenerative disorders, FER attainment has been well-described in frontotemporal dementia due to lesions located in the brain areas underpinning the processing of variant face features [8 –10]. Yet, the characteristics of SC deficit remain unclear in AD at early stages, despite being characterized by a rapid extension of neuropathological lesions from the entorhinal cortex and the hippocampal region to the neighboring temporal areas [11, 12] involved in the social brain network [13].

As already mentioned, the STS is also involved in processing gaze cues, notably determining eye-gaze direction. Identifying in which direction others are looking is a crucial step in social communication [14], presumably necessary to emotionally interact with others. According to Bediou et al. [15], gaze-direction processing is relatively unspoiled in early AD but recently, Insch et al. [16] showed that AD patients were less efficient than controls during a specific gaze-processing task.

To date, several publications have highlighted that patients suffering from AD experienced an increasing deficit in decoding facial emotions expression during the course of dementia [17 –19]. Possibly due to the lack of sufficient statistical power, current literature was conflicting concerning the features of the FER impairment in early stages AD. Weiss [19] found that patients with amnestic mild cognitive impairment (aMCI) performed worse than healthy controls (HC) recognizing fear and sadness. Meta-analysis findings have shown that patients with MCI significantly underperformed in FER of fear, anger, and sadness, especially for low-intensity emotions [20] when compared to healthy controls. Two longitudinal studies found a decline in FER parallel to the progression of AD [21]. Bediou et al. [15] showed overall FER impairment but preserved sex detection and eye-gaze processing ability in aMCI and moderate AD patients. Several studies found that happiness is globally the only emotion whose recognition remains preserved in AD early stages [22, 23]. On the other hand, disgust was the only emotion preserved in some large samples of patients with mild AD [18, 24]. As noted by a recent systematic review [25], most studies have used non-ecological tools such as photographs depicting only prototypical emotions which does not correspond to ordinary modes of communication and therefore may fail to characterize the subtle impairment of SC in early-stages AD. To deepen this disparate knowledge, we first focused on characterizing the FER deficit in a significant number of patients with mild AD by comparing them to a healthy control group using a FER task with different emotions intensities.

The aim of this study was to provide a better comprehension of FER impairment in the beginning stages of AD. We conducted socio-cognitive neuropsychological testing in participants diagnosed with AD at prodromal or mild dementia stages, and age-matched controls. Patients and controls underwent a task to gauge their ability to discriminate facial emotions (fear, anger, happiness, and disgust) expressed at different intensities. Among the 180 patients, 172 underwent a task to measure the ability to process eye-gaze direction to assess the potential deficit of this core component of social cognition [26] that appears crucial to contextualize facial emotion [27]. To ensure the integrity of invariable facial features pathway, we also compared sex recognition ability amongst 114 of the patients. Due to the possible relative preservation of the FFA in mild AD and the low-cognitive aspect of this task, we postulated that patients and healthy controls would have similar capacities concerning sex identification. Considering the results given by Bediou et al. [15] and the task’s low cognitive load, we predicted that AD patients should be capable to determine eye-gaze direction. We hypothesized that AD patients would show impairment in FER compared to controls. We also assumed that the depth of this deficit would be more pronounced for negative emotions and low-expressed emotions.

METHODS

Study design

The cross-sectional analyses were based on data drawn from the French multicentric prospective PACO (“Personnalité Alzheimer Comportement”) cohort. Patients with prodromal or mild AD were recruited from memory clinics during ambulatory visits. At baseline, study participants were provided with a psychiatric examination supplemented by an MRI, and a set of neuropsychological tasks comprising but not limited to social cognition, executive functions, and mood evaluation. We excluded patients with major visuo-perceptual or visuo-spatial deficits. The whole protocol was detailed in a previous paper [28]. Among the initial 237 participants, only those over 65 and with available data on FER were included in the current analysis (Fig. 1).

Fig. 1

The selecting procedure of the study sample, stem from the PACO study. Created with app.diagram.net.

Population description

Patients included in the PACO study met the following criteria: established diagnosis of mild stage AD according to NINCDS ADRDA standards [29] or prodromal stage [30] and characterized by the Clinical Dementia Rating [31] with CDR 0.5 at prodromal stage (considered aMCI) and CDR 1 at mild stage; over 50 years old; able to undergo clinical examination and complete neuropsychological testing; presence of a caregiver capable of informing the investigator about the onset of neuropsychiatric symptoms; and a Mini-Mental State Examination (MMSE) score ≥20.

The exclusion criteria included: patients with poorly managed and/or progressive psychiatric comorbidity (except for stabilized depression); patients receiving any psychotropic medication, except for antidepressant, hypnotic, anxiolytic, acetylcholinesterase inhibitors or memantine if prescribed and stabilized for more than 3 months; patients suffering from severe or unstable somatic pathologies; presence of NPS (except from mild depressed mood, anxiety, apathy, eating, or sleep disorders).

Control group

To confirm the presence of a deficit in FER in AD participants, we performed a comparison with a control group of healthy subjects initially recruited in parallel with the PACO study. To ensure representativeness, part of the control subjects were city-dwellers recruited through a regional seniors’ association and another part were from a rural caregiver support mission. Control subjects were also clinically examined and completed the neuropsychological test set. Only control subjects with available SC data and aged 65 years or older were included (Fig. 1).

Social cognition assessments

Participants underwent 3 social cognition tasks:

Sex recognition task

To ensure their ability to recognize invariable facial features and to perform a low-cognitive load task, subjects were required to complete a task based on sex recognition. Faces of 8 females and 8 males were morphed with a neutral face (computerized by averaging faces of 20 females and 20 males) to create a sexual continuum. None of the faces expressed any emotion. A total of 80 faces were randomly presented with 20% steps to the participants (16 faces for each step, i.e., 20%; 40%; 60%; 80%; 100%). Each face presentation was arranged as follows: fixation cross (0.5 s), face (1 s), and forced choice between male and female (until response).

Gaze direction task

A test to assess the capacity to detect eye gaze direction was performed. 64 photographs were created using 8 female and 8 male faces, looking straight ahead (0°) or looking right or left by 5°, 10°, or 15° of visual angle. Photographs were then randomly presented to participants for one second. Participants had to determine if the gaze was pointed towards them through a forced-choice answer (YES or NO). There was no response time limit.

Facial emotion recognition task

FER was assessed through a task initially designed in a previous study [25]. This computerized neuropsychological test was created to evaluate the recognition of fear, anger, disgust, happiness, and neutral face. An expression continuum ranging from 0% (neutral face) toward 100% (prototypical emotion) has been designed for each emotion by morphing the static color photographs of 2 males and 2 females depicting the facial emotion with a neutral face. A total of 96 photographs (i.e., 16 neutral face photographs and 4×20 photographs for each emotion) depicting each emotion by 20% steps were randomly presented to participants. After displaying a fixation cross for 0.5 s, a visual stimulus was presented for a duration of 1 s on a computer screen. Then patients had to select a forced choice label (neutral, fear, anger, disgust, happiness) which best described the emotional expression they saw. There was no response time limit.

Happiness was the only positive emotion while we grouped anger, disgust, and fear in the negative emotions for analysis purpose. The term gradient refers to emotional intensities ranging from 20% to 80% while the term prototypical refers only to 100% emotional intensity.

Statistical analysis

Controls and AD participants were compared using Student t tests for normally distributed data, or Kruskal-Wallis test in case of nonnormality and χ² tests. The multivariate associations between AD status and FER were investigated using a generalized linear model assuming that count data were Poisson distributed. Dispersion was systematically checked, and quasi-Poisson model were performed in case of overdispersion. Gaze direction and sex recognition tasks were investigated among the groups using linear models. Non-linearity, dependence of errors and multicollinearity were checked first. For all analyses, a p value less of 0.05 was considered statistically significant. Analyses were performed using R software, version 4.0.3 (2020-10-10 – “Bunny-Wunnies Freak out” – The R Foundation).

Covariates

Educational level, sex, age, and level of cognitive functioning reflected by the MMSE score were collected at baseline (M0) and then used as adjusting covariates in the analyses. Educational level ranged from 1 to 4: <5 years of education; 5 to 8 years; 9 to 11 years; >11 years, respectively.

RESULTS

Demographic data

Demographics and clinical data of patients with mild AD and healthy controls are shown in Table 1. Patients with AD were older than controls (p < 0.001^***) and had a lower education level (p < 0.001^***). No group differences were found on sex distribution (p < 0.474).

Table 1

Descriptive characteristics of the participants, PACO study and control group

	AD patients (n = 180)		Controls (n = 74)		p
	N	%	N	%
Sex					0.4740
Men	94	52.2	43	58.1
Women	86	47.8	31	41.9
Socio-cultural levels					<0.0001
1	7	4.4	2	2.7
2	50	31	7	9.5
3	66	41	22	29.7
4	38	23.6	43	58.1
	Mean	SD	Mean	SD
Age (y)	79.8	5.6	73.8	6.5	<0.001
MMSE	24.4	2.3	28.5	1.5	<0.001
Psychotropic medication	Yes	No
SSRI	35	145
SNRI	4	176
Atypical antidepressant	8	172
Neuroleptic	1	179
Benzodiazepine and Z-Drug	38	142
AChEIs	78	102

Sex determination task

As mentioned in Fig. 1, sex determination task data were available for 114 patients and 72 HC. Kruskal-Wallis analysis showed no difference between the groups in the distribution of sex task scores at all intensities (all p > 0.05) and for sex determination global score (p = 0.0536). Sex task’s performance are shown in Fig. 2.

Fig. 2

Sex recognition mean performance across the various intensities of sex morphs in AD group and control group.

Gaze direction task

As mentioned in Fig. 1, gaze direction task data were available for 172 patients and 67 HC. Eye-gaze eccentricity’s degree Kruskal-Wallis analysis showed distribution difference between the two groups at all gaze angles (at 5°, p < 0.001^***; at 10°, p = 0.009^**; at 15°, p < 0.001^***) except 0°, which corresponds to a neutral face. Figure 3 shows performance of the two groups on the gaze direction task.

Fig. 3

Gaze direction determination mean performance in the studied groups across the various degrees of eye-gaze eccentricity.

Univariate linear regression analysis showed lower performance in AD at 5° (p < 0.0001^***), 10° (p = 0.0008^***), and 15° (p = 0.0212^*). We then performed a multivariate analysis on age, sex, and educational level (Table 2).

Table 2

Eye-gaze direction determination comparison between Alzheimer’s disease group and control group, with linear regression model

	β	SE	T² statistic	p
Eccentricity 0°	0.0019	0.016	0.124	0.9020
Eccentricity 5°	–0.1248	0.039	–3.125	0.0020^**
Eccentricity 10°	–0.0551	0.024	–2.229	0.0268^*
Eccentricity 15°	–0.0150	0.009	–1.587	0.1139

The term eccentricity refers to the degree of deviation of the gaze. β, regression coefficient; SE, standard error; T²-statistic, multivariate T²-distribution.

Emotion recognition task

Figure 4 shows the results on the FER task for the different emotions and intensities of the AD group and control group.

Fig. 4

Emotion recognition mean performance (y-axis, number correctly identified emotions, max = 4) of the studied groups across the various emotional intensities.

Considering total emotions scores, univariate analysis showed that, compared to controls, AD participants had a lower ability to recognize emotions (Table 3). Negative emotions were not as well recognized as positive ones, while disgust appeared to be the emotion whose recognition was most discriminating between groups. Participants with AD consistently performed worse than controls on low-expressed emotions, yet there was no significant difference on prototypical emotions.

Table 3

Facial emotion recognition comparison between Alzheimer’s disease group and control group, with generalized linear regression model using poisson distribution

	β	SE	Z-statistic	p
All emotions	–0.107	0.018	–5.845	<0.0001^***
Positive emotion	–0.084	0.034	–2.458	0.0140^*
Negative emotions	–0.117	0.021	–5.368	<0.0001^***
Total Happiness	–0.084	0.034	–2.458	0.0140^*
Prototypical Happiness	–0.009	0.069	–0.139	0.8890
Low-expressed Happiness	–0.109	0.039	–2.257	0.0058^**
Total Fear	–0.115	0.037	–3.055	0.0022^**
Prototypical Fear	–0.125	0.073	–1.700	0.0891
Low-expressed Fear	–0.112	0.044	–2.543	0.0110^*
Total Disgust	–0.121	0.038	–3.167	0.0015^**
Prototypical Disgust	–0.109	0.075	–1.451	0.1470
Low-expressed Disgust	–0.126	0.045	–2.821	0.0047^**
Total Anger	–0.115	0.037	–3.077	0.0020^**
Prototypical Anger	–0.088	0.075	–1.160	0.2460
Low-expressed Anger	–0.125	0.043	–2.882	0.0039^**

Prototypical emotion refers only to the 100% expressed emotion while the term degraded emotion includes intensities of 20%, 40%, 60% and 80%. β, regression coefficient; SE, standard error; Z-statistic, Wald test.

Multivariate regression analysis adjusted on age, sex, and educational level showed no significant difference regarding prototypical emotions. AD patients performed worse on recognition of degraded fear (β= –0.167; Z = –2.765; p = 0.005^**) and degraded disgust (β= –0.1444; Z = –2.359; p = 0.018^*). The AD group tended to acknowledge degraded happiness worse than controls (β= –0.092; Z = –1.705; p = 0.088). There were no significant differences in anger recognition, either degraded or prototypical. Multivariate analyses after exclusion of psychotropics consumers showed that AD patients performed worse on recognition of degraded disgust (β= –0.147; Z = –2.501; p = 0.012^*). The AD group tended to acknowledge degraded fear (β= –0.106; Z = –1.849; p = 0.064) and degraded happiness worse than controls (β= –0.071; Z = –1.364; p = 0.173). There were no significant differences in anger recognition, either degraded or prototypical. Analyses were also performed by adjusting for overall cognitive level (reflected by the MMSE-score) with only degraded disgust recognition showing a significant difference between the two groups (β= –0.177; Z = –2.595; p = 0.009^**). Data are shown in the Supplementary Material.

DISCUSSION

The goal of this research was to explore FER and GD identification in prodromal and mild AD compared to healthy controls. To our knowledge, emotion recognition of computerized morphed facial expressions has never been examined in such a large population of AD patients. Our cross-sectional study was based on the comparison of 180 mild AD patients with a control group of 74 healthy age-matched controls to lessen the possible effect of aging on emotion recognition capacities [32]. To ensure that FER was unrelated to a major visuo-perceptual deficit, we first showed that, in accordance with our hypothesis, the two groups exhibited no difference in sex determination. Contrary to our assumption, we found that AD patients were impaired at determining GD at 5° and 10° of eccentricity. These results suggest that for a straight-forward gaze or for a high degree of eccentricity patients could still be able to determine whether they are being looked at, or not, but that their performance is more attenuated for smaller deviations. Using a similar protocol but with a sample of only 10 AD’s patients, Bediou [15] found no impairment of GD determination while Insch [16] found a significant effect of AD on GD. Unadjusted analyses showed that AD patients performed globally worse than controls in FER, especially in regard to negative emotions. Prototypical emotion recognition was preserved in both groups while low-expressed emotion recognition was impaired in the AD group suggesting that the level of emotional intensity could mitigate the depth of the FER deficit. Then, linear regression analysis showed that the AD group had difficulties recognizing low-expressed fear and disgust. The AD group also showed a tendency to fail low-expressed happiness recognition. In contrast, prototypical emotion recognition was comparable between the two groups. Our results are in partial contradiction to those of Spoletini [18] and Henry [24] who confirmed a more pronounced impairment of emotional recognition for weakly expressed emotions but revealed no impairment of disgust recognition in early AD. In line with Granato [33] and Weiss [19], our results suggest that anger recognition is preserved in mild AD.

According to Haxby’s model [34], the FFA, located in the inferior temporal cortex, is largely responsible for the perception of invariant elements of the face, notably sex and identity. The later involvement of this region in AD and the relative simplicity of this neurocognitive task may explain why sex determination is preserved in our sample. In contrast, FER and gaze processing rely on structures that are affected earlier by AD’s neuropathological course such as the STS [35] and the amygdala [36, 37]. Although the amygdala is allegedly involved in a wide range of facial emotional expressions [38], it appears to play a more pivotal function in fear perception [39]. The insula, specifically the ventral anterior insula, has been classically associated with recognition and behavioral responses to disgust [40, 41]. Albeit being purely speculative in the absence of any imaging evidence, our theory is that the rapid extension of neurofibrillary tangles toward the amygdala and the insula in early stages of AD [11] could explain the more pronounced deficit found in our research for low-expressed fear and disgust. Anger recognition is thought to rely more specifically on the anterior cingulate cortex, the ventral striatum, and the orbitofrontal cortex [42], affected regions in later stages of AD. Happiness recognition seems to depend on the anterior cingulate cortex but also in large part on the amygdala [38], potentially explaining the trend found in our study.

The strengths of this study include the large number of participants and their rigorous inclusion based on international diagnostic criteria for dementia and AD. Additionally, whereas previous research often included patients with several dementia etiologies at various stages, our study focused on homogenous clinical samples, including patients with probable AD at prodromal or mild stage. The cognitive state of participants was relatively high (mean MMSE = 24.4) allowing for an early assessment of FER and thus facilitating extrapolation to practical clinical use. The protocol of the FER task based on the creation of gradual intensities through face morphing is close to real-life situations where emotions are rarely expressed in a prototypical way. Using different intensities permitted the study to bypass the ceiling effect frequently observed concerning happiness, due to its relatively simple recognition for both elderly controls and patients.

Several limitations must be taken into consideration when interpreting these results. Sadness and surprise have been discarded from the social cognition task we used because of their low discriminatory potential leading to the possible artifact of studying only one positive emotion versus three negatives ones. Apart from the ceiling effect found in happiness recognition, due to the wide use of only prototypical emotion recognition tasks, the relative preservation of happiness recognition in many studies could possibly be explained by participants’ tendency to choose happiness in moments of hesitation. Due to low acceptability of cerebrospinal fluid intake by lumbar puncture for research purposes and to ensure having a large pool of participants, we based the diagnosis on widely used international clinical criteria and MRI brain imaging. The hindrance posed by the lack of standardized tests and ecological instruments to consistently investigate the ability of AD patients to recognize facial emotions may explain the somewhat disparity in results of the current FER literature. The self-design FER task we used has not been validated by a dedicated study and was initially created for schizophrenia [43] but was notably similar to the Emotion Recognition Test (ERT), a validated task which has the already evoked interest of presenting degraded emotions [44]. The design of a dynamic emotion recognition task, for example, using short video clips as Kumfor et al. did [21], would allow a better match with current knowledge about the redundant functions of the two neural pathways involved in FER [7]. In fact, intonation, prosody, or gestures are all crucial social cues that we consider concomitantly when analyzing facial emotions in real life situations accounting for the ecological defect of tasks exploring only one of these aspects at a time. Thus, Henry [45] showed that emotional recognition could be normalized in MCI patients compared to HC using dynamic emotion recognition tasks. Lacking data in the control group, we could not use the Geriatric Depression Scale (GDS) as a covariate in order to reduce a possible vulnerability for negative emotions [46] when depressive symptoms co-occurs.

As described in previous studies [47, 48], it appears risky to completely rule out the hypothesis that a dysexecutive impairment in AD could be partly responsible for FER alteration especially when using static faces. Moreover, the literature disparity in FER results may be due to the variability between studies of participants global cognitive impairment. Indeed, the MMSE score is only a reflection of a global, heterogeneous deflection and may lack consistency in assessing the neuronal degradation of the emotion recognition network. A more pertinent indicator of the neuropathological lesions extent from the hippocampus to the neighboring temporal regions could be the use of the Free and Cued Selective Reminding Test as an adjustment variable [49]. However, in the present study, the large number of photographs presented to each participant to ensure instructions’ proper understanding and patients relatively high overall cognitive level may help to mitigate this bias. Some studies have suggested that FER relies more specifically on the ocular region [50] and thus depends on visual scanning patterns [23]. Patients with AD may show more visual wandering and a certain inefficiency to consider several facial cues simultaneously compared to healthy subjects of the same age, impacting their ability to recognize emotions [23]. Visual wandering could depend on an impairment of executive functions [51] but also on specifically altered visual processing strategies [52]. The effect of the aging process on executive functions could also be responsible for a modification of visual scanning strategies of facial emotion [53]. A meta-analysis concluded that aging negatively affected emotion recognition, not including disgust [32]. While adjusting on the deterioration of the executive functions and the modifications of the eye-tracking skills, some authors found only a marked impairment of fear recognition thus implying that fear could be a more archaic emotion, of predominantly amygdalian processing, in order to keep its role of physiological alarm in spite of cognitive damage [54].

Considering the multiple discrepancies in the current literature, we can also question the relevance of a socio-cognitive phenotyping of early AD. Conversely, the ability for AD patients to recognize facial emotions would rather depend on their neuropsychological profile [55], in the sense of a specific subset of impaired/preserved cognitive functions for each person, and on a sum of influencing factors such as density of a social network, depression history [56], reduced visual acuity or hearing loss. Indeed, the influence of social isolation on the onset of neurocognitive disorders has been highlighted by recent studies [57, 58], thus questioning the preservation of social cognitive abilities in the absence of regular stimulation.

As previously explained, only studies using standardized tests of emotion recognition in the most ecological conditions possible will we be able to accurately assess the characteristics of the FER impairment in AD. Longitudinal follow-up of the deterioration of facial emotion recognition abilities during AD’s course may facilitate our insight into the determinants and consequences of this socio-emotional deficiency. Furthermore, our understanding of the neural networks supposedly involved in behavioral and emotional changes that appear in early AD would be enhanced by studies combining neuroimaging and neuropsychological assessments. The current study’s findings suggest an early impairment in detecting the GD of others and in the recognition of certain facial emotions in AD. Further analysis would be necessary to evaluate the clinical impact of these socio-emotional manifestations and the role they may play in the neuropsychiatric symptoms’ genesis. The impairment of FER in AD patients could contribute to the caregiver’s burden and depressive symptomatology [59], especially since caregivers seem to be insufficiently aware of this socio-cognitive deficit [60]. In order to limit a daily clinical impact, specific socio-emotional remediation programs [61] could improve the social interactions efficiency in patients in early stages of AD. Cognitive remediation could even be used to enhance the quality of visual scanning of the face to more easily recognize facial emotions by focusing on eye-tracking and limiting visual wandering [62]. In order to break the pattern of defective social interactions that promote behavioral and psychological symptoms, caregivers could also benefit from therapeutic education in order to adjust their attitude and communication style [63] with subjects presenting deficient eye-gaze processing or a lack of emotion recognition. Finally, the clinical implications of our study could sensitize socially competent healthcare teams to the importance of providing an empathetic and adapted approach to the impairment of social cognition in AD.

Footnotes

ACKNOWLEDGMENTS

The PACO study was funded by the National French Program of Hospital Clinical Research (Programme Hospitalier de Recherche Clinique National 2009 and 2010), France Alzheimer, Janssen laboratory grants.

Authors’ disclosures available online ().

The supplementary material is available in the electronic version of this article: .

References

Adolphs

(2009) The social brain: Neural basis of social knowledge. Annu Rev Psychol 60, 693–716.

Hugenberg

, Wilson

(2013) Faces are central to social cognition. Handbook Social Cogn. pp. 167–193.

Dalgleish

(2004) The emotional brain. Nat Rev Neurosci 5, 583–589.

Kohler

, Turner

, Bilker

, Brensinger

, Siegel

, Kanes

, Gur

(2003) Facial emotion recognition in schizophrenia: Intensity effects and error pattern. Am J Psychiatry 160, 1768–1774.

Harms

, Martin

, Wallace

(2010) Facial emotion recognition in autism spectrum disorders: A review of behavioral and neuroimaging studies. Neuropsychol Rev 20, 290–322.

Elamin

, Pender

, Hardiman

, Abrahams

(2012) Social cognition in neurodegenerative disorders: A systematic review. J Neurol Neurosurg Psychiatry 83, 1071–1079.

Duchaine

, Yovel

(2015) A revised neural framework for face processing. Annu Rev Vis Sci 1, 393–416.

Bediou

, Brunelin

, d’Amato

, Fecteau

, Saoud

, Hénaff

M-A

, Krolak-Salmon

(2012) A comparison of facial emotion processing in neurological and psychiatric conditions. Front Psychology 3, 98.

Bertoux

, de Souza

, O’Callaghan

, Greve

, Sarazin

, Dubois

, Hornberger

(2015) Social cognition deficits: The key to discriminate behavioral variant frontotemporal dementia from Alzheimer’s disease regardless of amnesia? J Alzheimers Dis 49, 1065–1074.

10.

Bora

, Velakoulis

, Walterfang

(2016) Meta-analysis of facial emotion recognition in behavioral variant frontotemporal dementia: Comparison with Alzheimer disease and healthy controls. J Geriatr Psychiatry Neurol 29, 205–211.

11.

Braak

, Braak

(1991) Neuropathological stageing of Alzheimer-related changes. Acta Neuropathol 82, 239–259.

12.

Chetelat

, Baron

J-C

(2003) Early diagnosis of Alzheimer’s disease: Contribution of structural neuroimaging. Neuroimage 18, 525–541.

13.

Billeke

, Aboitiz

(2013) Social cognition in schizophrenia: From social stimuli processing to social engagement. Front Psychiatry 4, 4.

14.

Kleinke

(1986) Gaze and eye contact: A research review. Psychol Bull 100, 78–100.

15.

Bediou

, Ryff

, Mercier

, Milliery

, Hénaff

M-A

, D’Amato

, Bonnefoy

, Vighetto

, Krolak-Salmon

(2009) Impaired social cognition in mild Alzheimer disease. J Geriatr Psychiatry Neurol 22, 130–140.

16.

Insch

, Slessor

, Warrington

, Phillips

(2017) Gaze detection and gaze cuing in Alzheimer’s disease. Brain Cogn 116, 47–53.

17.

Lavenu

, Pasquier

(2005) Perception of emotion on faces in frontotemporal dementia and Alzheimer’s disease: A longitudinal study. Dement Geriatr Cogn Disord 19, 37–41.

18.

Spoletini

, Marra

, Iulio

, Gianni

, Sancesario

, Giubilei

, Trequattrini

, Bria

, Caltagirone

, Spalletta

(2008) Facial emotion recognition deficit in amnestic mild cognitive impairment and Alzheimer disease. Am J Geriatr Psychiatry 16, 389–398.

19.

Weiss

, Kohler

, Vonbank

, Stadelmann

, Kemmler

, Hinterhuber

, Marksteiner

(2008) Impairment in emotion recognition abilities in patients with mild cognitive impairment, early and moderate Alzheimer disease compared with healthy comparison subjects. Am J Geriatr Psychiatry 16, 974–980.

20.

Bora

, Yener

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

21.

Kumfor

, Irish

, Leyton

, Miller

, Lah

, Devenney

, Hodges

, Piguet

(2014) Tracking the progression of social cognition in neurodegenerative disorders. J Neurol Neurosurg Psychiatry 85, 1076–1083.

22.

Park

, Kim

, Shin

, Kim

, Sohn

, Park

H-J

, Youn

J-H

, Lee

J-Y

(2017) Behavioral and neuroimaging evidence for facial emotion recognition in elderly korean adults with mild cognitive impairment, Alzheimer’s disease, and frontotemporal dementia. Front Aging Neurosci 9, 389.

23.

Hot

, Klein-Koerkamp

, Borg

, Richard-Mornas

, Zsoldos

, Paignon Adeline

, Thomas Antérion

, Baciu

(2013) Fear recognition impairment in early-stage Alzheimer’s disease: When focusing on the eyes region improves performance. Brain Cogn 82, 25–34.

24.

Henry

, Ruffman

, McDonald

, O’Leary

M-AP

, Phillips

, Brodaty

, Rendell

(2008) Recognition of disgust is selectively preserved in Alzheimer’s disease. Neuropsychologia 46, 1363–1370.

25.

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.

26.

Itier

, Batty

(2009) Neural bases of eye and gaze processing: The core of social cognition. Neurosci Biobehav Rev 33, 843–863.

27.

Sander

, Grandjean

, Kaiser

, Wehrle

, Scherer

(2007) Interaction effects of perceived gaze direction and dynamic facial expression: Evidence for appraisal theories of emotion. Eur J Cogn Psychol 19, 470–480.

28.

the PACO group, Rouch

, Dorey

J-M

, Boublay

, Henaff

M-A

, Dibie-Racoupeau

, Makaroff

, Harston

, Benoit

, Barrellon

M-O

, Fédérico

, Laurent

, Padovan

, Krolak-Salmon

(2014) Personality, Alzheimer’s disease and behavioural and cognitive symptoms of dementia: The PACO prospective cohort study protocol. BMC Geriatr 14, 110.

29.

McKhann

, Drachman

, Folstein

, Katzman

, Price

, Stadlan

(1984) Clinical diagnosis of Alzheimer’s disease: Report of the NINCDS-ADRDA Work Group under the auspices of Department of Health and Human Services Task Force on Alzheimer’s Disease. Neurology 34, 939–944.

30.

Albert

(2011) Changes in cognition. Neurobiol Aging 32, S58–S63.

31.

Morris

(1997) Clinical dementia rating: A reliable and valid diagnostic and staging measure for dementia of the Alzheimer type. Int Psychogeriatr 9(Suppl 1), 173–176; discussion 177-178.

32.

Ruffman

, Henry

, Livingstone

, Phillips

(2008) A meta-analytic review of emotion recognition and aging: Implications for neuropsychological models of aging. Neurosci Biobehav Rev 32, 863–881.

33.

Granato

, Vinekar

, Godefroy

, Gansberghe

J-PV

, Bruyer

(2020) The ability to visually recognize facial emotions during the initial stages of Alzheimer’s. Open J Psychiatry 10, 187–204.

34.

Haxby

, Gobbini

(2011), Distributed Neural Systems for Face Perception, Oxford University Press.

35.

Carlin

, Calder

(2013) The neural basis of eye gaze processing. Curr Opin Neurobiol 23, 450–455.

36.

Sapey-Triomphe

L-A

, Heckemann

, Boublay

, Dorey

J-M

, Hénaff

M-A

, Rouch

, Padovan

, Hammers

, Krolak-Salmon

, Alzheimer’s Disease Neuroimaging Initiative (2015) Neuroanatomical correlates of recognizing face expressions in mild stages of Alzheimer’s disease. PLoS One 10, e0143586.

37.

Kawashima

, Sugiura

, Kato

, Nakamura

, Hatano

, Ito

, Fukuda

, Kojima

, Nakamura

(1999) The human amygdala plays an important role in gaze monitoring. A PET study. Brain 122(Pt 4), 779–783.

38.

Fusar-Poli

, Placentino

, Carletti

, Landi

, Allen

, Surguladze

, Benedetti

, Abbamonte

, Gasparotti

, Barale

, Perez

, McGuire

, Politi

(2009) Functional atlas of emotional faces processing: A voxel-based meta-analysis of 105 functional magnetic resonance imaging studies. J Psychiatry Neurosci 34, 418–432.

39.

Adolphs

(2008) Fear, faces, and the human amygdala. Curr Opin Neurobiol 18, 166–172.

40.

Krolak-Salmon

, Hénaff

M-A

, Isnard

, Tallon-Baudry

, Guénot

, Vighetto

, Bertrand

, Mauguière

(2003) An attention modulated response to disgust in human ventral anterior insula. Ann Neurol 53, 446–453.

41.

Woolley

, Strobl

, Sturm

, Shany-Ur

, Poorzand

, Grossman

, Nguyen

, Eckart

, Levenson

, Seeley

, Miller

, Rankin

(2015) Impaired recognition and regulation of disgust is associated with distinct but partially overlapping patterns of decreased gray matter volume in the ventroanterior insula. Biol Psychiatry 78, 505–514.

42.

Blair

RJR

, Morris

, Frith

, Perrett

, Dolan

(1999) Dissociable neural responses to facial expressions of sadness and anger. Brain 122, 883–893.

43.

Bediou

, Krolak-Salmon

, Saoud

, Henaff

M-A

, Burt

, Dalery

, D’Amato

(2005) Facial expression and sex recognition in schizophrenia and depression. Can J Psychiatry 50, 525–533.

44.

Kessels

RPC

, Montagne

, Hendriks

, Perrett

, de Haan

EHF

(2014) Assessment of perception of morphed facial expressions using the Emotion Recognition Task: Normative data from healthy participants aged 8-75. J Neuropsychol 8, 75–93.

45.

Henry

, Thompson

, Rendell

, Phillips

, Carbert

, Sachdev

, Brodaty

(2012) Perception of biological motion and emotion in mild cognitive impairment and dementia. J Int Neuropsychol Soc 18, 866–873.

46.

Leppänen

(2006) Emotional information processing in mood disorders: A review of behavioral and neuroimaging findings. Curr Opin Psychiatry 19, 34–39.

47.

Cadieux

, Greve

(1997) Emotion processing in Alzheimer’s disease. J Int Neuropsychol Soc 3, 411–419.

48.

Hargrave

, Maddock

, Stone

(2002) Impaired recognition of facial expressions of emotion in Alzheimer’s disease. J Neurophysiol 14, 64–71.

49.

Philippi

, Noblet

, Duron

, Cretin

, Boully

, Wisniewski

, Seux

, Martin-Hunyadi

, Chaussade

, Demuynck

, Kremer

, Lehéricy

, Gounot

, Armspach

, Hanon

, Blanc

(2016) Exploring anterograde memory: A volumetric MRI study in patients with mild cognitive impairment. Alzheimers Res Ther 8, 26.

50.

Polet

, Hesse

, Kullmann

, Morisot

, Louchart de la Chapelle

, Pesce

(2017) Analyse de l’exploration visuelle par électro-oculographie motrice en tâche de cognition sociale chez des patients atteints de pathologies neurodégénératives. NPG Neurol Psychiatr Gériatr 17, 188–195.

51.

García-Rodríguez

, Vincent

, Casares-Guillén

, Ellgring

, Frank

(2012) The effects of different attentional demands in the identification of emotional facial expressions in Alzheimer’s disease. Am J Alzheimers Dis Other Demen 27, 530–536.

52.

Ogrocki

, Hills

, Strauss

(2000) Visual exploration of facial emotion by healthy older adults and patients with Alzheimer disease. Neuropsychiatry Neuropsychol Behav Neurol 13, 271–278.

53.

Circelli

, Clark

, Cronin-Golomb

(2013) Visual scanning patterns and executive function in relation to facial emotion recognition in aging. Neuropsychol Dev Cogn B Aging Neuropsychol Cogn 20, 148.

54.

Abbruzzese

, Magnani

, Robertson

, Mancuso

(2019) Age and gender differences in emotion recognition. Front Psychol 10, 2371.

55.

Luzzi

, Piccirilli

, Provinciali

(2007) Perception of emotions on happy/sad chimeric faces in Alzheimer disease: Relationship with cognitive functions. Alzheimer Dis Assoc Disord 21, 130–135.

56.

Anderson

, Shippen

, Juhasz

, Chase

, Thomas

, Downey

, Toth

, Lloyd-Williams

, Elliott

, Deakin

(2011) State-dependent alteration in face emotion recognition in depression. Br J Psychiatry 198, 302–308.

57.

Sommerlad

, Sabia

, Singh-Manoux

, Lewis

, Livingston

(2019) Association of social contact with dementia and cognition: 28-year follow-up of the Whitehall II cohort study. PLoS Med 16, e1002862.

58.

Livingston

, Huntley

, Sommerlad

, Ames

, Ballard

, Banerjee

, Brayne

, Burns

, Cohen-Mansfield

, Cooper

, Costafreda

, Dias

, Fox

, Gitlin

, Howard

, Kales

, Kivimäki

, Larson

, Ogunniyi

, Orgeta

, Ritchie

, Rockwood

, Sampson

, Samus

, Schneider

, Selbæk

, Teri

, Mukadam

(2020) Dementia prevention, intervention, and care: 2020 report of the Lancet Commission. Lancet 396, 413–446.

59.

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.

60.

Spitzer

, Shafir

, Lerman

, Werner

(2019) The relationship between caregiver burden and emotion recognition deficits in persons with MCI and early AD: The mediating role of caregivers’ subjective evaluations. Alzheimer Dis Assoc Disord 33, 266–271.

61.

García-Casal

, Goñi-Imizcoz

, Perea-Bartolomé

, Soto-Pérez

, Smith

, Calvo-Simal

, Franco-Martín

(2017) The efficacy of emotion recognition rehabilitation for people with Alzheimer’s disease. J Alzheimers Dis 57, 937–951.

62.

Polet

, Mirzai

, Hesse

, Morisot

, Kullmann

, Louchart de la Chapelle

, Pesce

, Iakimova

(2021) Remédiation de la reconnaissance des émotions faciales dans la maladie d’Alzheimer et effets sur les stratégies d’observation, les troubles du comportement et le fardeau de l’aidant [Rehabilitation of facial emotion recognition in Alzheimer’s disease and effects on observational strategy, behavioral disorders and family caregiver burden]. NPG Neurol Psychiatr G’riatr 21, 25–36.

63.

Schulz

, Martire

(2004) Family caregiving of persons with dementia: Prevalence, health effects, and support strategies. Am J Geriatr Psychiatry 12, 240–249.