Given the aging of the population, the prevalence of dementia such as Alzheimer’s disease (AD) is increasing. It is now largely accepted that early identification of AD is crucial. An important scientific and clinical challenge is to clarify the neuropsychological profile of individuals with mild cognitive impairment (MCI) to assist the identification of early AD-related cognitive deficits. The concept of MCI corresponds to the prodromal stage of AD in several, if not most, older adults (for reviews on MCI, see [
Among the neuropsychological characteristics of individuals with MCI, mentalizing deficits have already been documented [
Verbal irony comprehension is generally defined as a linguistic process used to express information which is directly, or indirectly, opposed to its literal interpretation [
The main objectives of this study are twofold: (a) to corroborate previous results indicating the presence of mentalizing deficits in older adults with MCI and (b) to examine the implication of mentalizing in the comprehension of verbal assertions based on mental state inferences (i.e., ironic and sincere assertions) in MCI. To this end, participants were administered two different tasks, one evaluating mentalizing capacities and the other verbal irony comprehension by means of short stories. The mentalizing task assessed the integrity of first- (i.e., inferring somebody else’s mental state) and second-order (i.e., inferring what somebody might think about another’s mental state) mentalizing, as well as nonsocial reasoning (i.e., inference of facts or events that implicate no human beings). The verbal irony task assessed the interpretation of ironic and sincere scenarios. Considering existing bodies of evidence of second-order mentalizing deficits in individuals with MCI [
Participants included 30 older adults with MCI (age range: 60–83 years) and 30 healthy control (HC) subjects (age range: 55–82 years). Participants with MCI were identified and referred to the research team by experienced geriatricians, neurologists, or general practitioners, or through advertisements displayed in different clinics. All MCI participants met the clinical core criteria of Albert et al. [
For all participants, exclusion criteria were as follows: (a) history of traumatic brain injury; (b) history of stroke or other cerebrovascular disorders; (c) history of delirium (in the last six months); (d) formal intracranial surgery; (e) history of neurological disorder of cerebral origin or associated with another dementia state (e.g., parkinsonism, multiple sclerosis, etc.); (f) history of encephalitis or bacterial meningitis; (g) unstable metabolic or medical condition (e.g., uncontrolled diabetes or hypothyroidism); (h) history or current diagnosis of a psychiatric illness or dementia according to the DSM-IV [
An informed consent form was signed by each participant at the beginning of the first assessment session. This study was approved by the Ethics Research Board of the
Participants were administered a comprehensive battery of clinical and neuropsychological tests in order to verify inclusion/exclusion criteria and to characterize their cognitive and clinical status. The battery included instruments to assess depressive symptoms (Geriatric Depression Scale [
First- and second-order mentalizing abilities were evaluated using the
The
The SSICT consists in written presentation of 14 short stories describing everyday life situations, each including four assertions (for a sample of the task see the following). Sample story in French adapted from Eviatar and Just [
Julie est bavarde et elle ne peut s’empêcher de tout commenter. ( Son ami Robert a l’habitude de l’entendre parler sans cesse et cela le fait bien rire. ( Un jour, ils se rencontrent pour déjeuner, mais Julie est fatiguée et elle ne parle presque pas. (
The first sentence always presents the general context of the scenario, whereas the second phrase describes the specific context. The third statement presents a triggering event. Finally, the fourth sentence is the one that has to be evaluated as being either (a) sincere, (b) ironic, or (c) descriptive. Of the 14 scenarios, five describe a sincere situation, five are ironic, and four are descriptive. The fourth statement of the
To correctly interpret both sincere and ironic scenarios, participants must infer or metarepresent mental states. It is assumed that the “mentalizing load” is comparable between sincere and ironic scenarios since, in both cases, the interaction between two protagonists is depicted. Moreover, for both sincere and ironic stories, participants have to infer a mental state in one protagonist by taking into account the context described in the first three literal assertions. The variable of interest in the SSICT is the correct interpretation (i.e., response accuracy) of the fourth sentence, which is scored as 1 or 0.
Participants were assessed individually over three sessions. The first session included the signature of the informed consent form and the administration of the clinical and neuropsychological tests or questionnaires; this session lasted approximately two hours. Mentalizing abilities (Combined Stories Test) and verbal irony comprehension (SSICT) were assessed during the second and third sessions, in a counterbalanced order. These sessions lasted approximately 30 to 45 minutes each.
Participants sat at about 1 m from the computer screen and were asked to read the stories out loud. When finished, the experimenter asked questions (first- and second-order mentalizing, nonsocial reasoning, and attention/memory) and wrote down the participant’s answer on the response sheet.
Participants sat at about 1 m from the computer screen. They were asked to read out loud the task instructions, which appeared on the computer screen. They were then asked to read the written stories out loud in a self-paced manner. When the first two sentences of a story were read, participants pressed the “Enter” button in order to see and read sentences 3 and 4 (the first two sentences remained on the screen at this stage). Following reading of the fourth sentence, subjects were asked whether the story was (1) sincere, (2) ironic, or (3) descriptive. They had to answer by pressing the corresponding number on a keypad placed in front of them. Stories were counterbalanced in a random order to control for a possible sequence effect.
Except for sex distribution in the groups, which was analysed using Pearson Chi-square test, Student’s
Exploratory analyses were first conducted to examine normal distribution assumptions required for inferential analyses. All variables of interest were normally distributed. Regarding the Combined Stories Test, proportions of correct responses were calculated and Student’s
Pearson’s correlations were used to assess the association between the key variables of the experimental tasks (i.e., stories evaluating second-order mentalizing in the Combined Stories Test and stories evaluating comprehension of ironic and sincere scenarios in the SSICT). With regard to the SSICT, the relationship between results for ironic and sincere stories was also analyzed, to examine statistically the extent to which these two conditions rely on similar cognitive processes. Moreover, Pearson’s correlations were performed in order to analyse associations between the score on variables of interest in the Combined Stories Test and the SSICT with measures of executive functions and episodic memory (mean score of immediate free and total recall from the Free Recall/Cued Recall 16-Item Test [RL/RI-16]). In order to limit the number of analyses, a composite score was computed for executive functions, including measures of manipulation of information in working memory (score on the WAIS-III letter-number sequencing task), controlled inhibition (time and total number of errors on the third condition of the D-KEFS Color-Word Interference Test), and mental flexibility (time and total number of errors on the fourth conditions of the D-KEFS Color-Word Interference and the Trail Making Tests).
For all analyses, an alpha level of .05 was used.
Table
Mean (SD) and significance levels of demographic and neuropsychological characteristics of participants.
Measures | MCI | HC |
|
|
---|---|---|---|---|
( |
( |
|
||
|
||||
Age (years) | 73.9 (6.1) | 71.9 (8.2) | 1.2 | 0.230 |
Sexa | 18F/12M | 19F/11M |
|
0.791 |
Education (years) | 13.6 (5.3) | 14.2 (4.0) | −0.5 | 0.618 |
|
||||
GDS (/30) | 6.8 (4.6) | 2.6 (2.6) | 4.3 |
0.000 |
|
||||
QPC (/10) | 3.6 (2.3) | 0.93 (1.4) | 5.5 |
0.000 |
|
||||
MoCA (/30) | 23.0 (3.2) | 26.5 (2.4) | −4.8 |
0.000 |
Dementia Rating Scale (/144) | 134.0 (5.7) | 139.8 (2.8) | −5.0 |
0.000 |
|
||||
ACDS-ADL inventory (/45) | 37.6 (5.6) | 41.9 (3.4) | −3.6 |
0.001 |
|
||||
Letter-number sequencing (WAIS-IV; scaled score) | 8.4 (2.0) | 9.7 (1.6) | −2.6 |
0.013 |
D-KEFS Color-Word Interference Test Condition 3 (sec) | 86.9 (26.7) | 64.3 (12.2) | 4.1 |
0.000 |
D-KEFS Color-Word Interference Test Condition 3 (errors) | 3.0 (2.7) | 1.1 (1.6) | 3.2 |
0.003 |
D-KEFS Color-Word Interference Test Condition 4 (sec) | 105.6 (39.4) | 72.4 (21.2) | 4.0 |
0.000 |
D-KEFS Color-Word Interference Test Condition 4 (errors) | 5.3 (5.0) | 1.9 (1.6) | 3.6 |
0.001 |
D-KEFS Trail Making Test Condition 4 (sec) | 156.4 (70.3) | 109.6 (33.6) | 3.2 |
0.002 |
D-KEFS Trail Making Test Condition 4 (errors) | 0.9 (2.9) | 0.0 (0.2) | 1.6 | 0.129 |
|
||||
Copy of Rey Figure (/36) | 28.9 (4.9) | 31.9 (3.2) | −2.7 |
0.009 |
|
||||
Letter (T-N-P) fluency (number of words in 1 min) | 29.2 (9.0) | 35.4 (10.1) | −2.5 |
0.015 |
Semantic fluency (number of words in 1 min) | 14.7 (4.9) | 17.3 (4.4) | −2.2 | 0.033 |
Boston Naming Test (/15) | 12.6 (1.7) | 13.4 (1.3) | −2.2 | 0.033 |
|
||||
Pyramids and Palm Trees Test (/52) | 48.5 (3.2) | 49.9 (1.8) | −2.1 | 0.041 |
|
||||
Recall of the Rey Figure (3 min; /36) | 12.1 (5.8) | 17.6 (6.0) | −3.6 |
0.001 |
RL-RI free recall (/16)b | 6.0 (1.8) | 9.8 (1.2) | −9.7 |
0.000 |
RL-RI total recall (/16)c | 12.2 (2.5) | 15.5 (0.7) | −7.1 |
0.000 |
RL-RI delayed free recall (/16) | 6.6 (3.7) | 12.6 (2.1) | −7.8 |
0.000 |
RL-RI total delayed total recall (/16) | 12.7 (2.9) | 15.7 (0.4) | −5.8 |
0.000 |
aStatistical differences for distribution of gender in each group were examined using Pearson Chi-Square test.
bThis score was calculated as the mean number of words retrieved over the three free recall trials.
cThis score was calculated as the mean total number of words retrieved on all free plus cued recall trials.
Two participants out of 30 were excluded from the MCI group because of missing data or withdrawal from the study. Thus, 28 MCI participants were included in the final analyses. Figure
Mean (±SEM) proportions of correct scores on second-order mentalizing, nonsocial reasoning, and first-order mentalizing of the Combined Stories Test. MCI = elderly persons with mild cognitive impairment; HC = healthy controls participants; SEM = standard error of the mean
Figure
Mean proportions of correct responses (±SEM) for ironic and sincere stories of the Short Scenario Irony Comprehension Task. MCI = elderly persons with mild cognitive impairment; HC = healthy control participants; SEM = standard error of the mean
Overall, the second-order mentalizing score of the Combined Stories Test was significantly related to the scores on both ironic (
With respect to the SSICT, significant positive relationships were found between score in the sincere stories and mean scores of free (
Finally, the second-order mentalizing score of the Combined Stories Test was significantly and positively related to the mean scores of free (
The main purposes of this study were (1) to confirm second-order mentalizing impairment in MCI and (2) to examine the relative implication of mental state inferences in the understanding of ironic and sincere assertions. With respect to second-order mentalizing abilities, results confirmed that MCI participants had significantly greater difficulties than their healthy counterparts. However, the groups were comparable regarding nonsocial reasoning and first-order mentalizing capacities. Regarding comprehension of ironic and sincere stories involving second-order mentalizing, participants with MCI were impaired relative to the HC group. That is, MCI participants had difficulty in correctly interpreting both ironic and sincere stories.
The present findings extend prior results published by Gaudreau et al. [
Results of the Combined Stories Test [
With respect to the SSICT, the MCI participants’ difficulty in interpreting ironic written scenarios is in accord with previous results showing impaired verbal irony comprehension in this population [
The significant relationship between second-order mentalizing and irony comprehension can also be explained by other factors. For instance, it could be that these cognitive processes are affected essentially to the same degree by the underlying neuropathology in MCI. Alternatively, the two measures could reflect to some extent a similar cognitive mechanism or process and thus impaired performance on one measure would be accompanied by impaired performance on the other. In fact, it is possible that the positive and significant association between performances in both experimental tasks derives from the underlying abilities in language and executive domains involved in both mental inferences abilities and verbal irony comprehension. Further studies will be necessary to verify these hypotheses.
The study of the relationships among the variables of interest brought interesting results, but the authors acknowledge that a better analysis strategy would have been to use linear regression instead of Pearson’s correlation. Unfortunately, the sample size in this study was not large enough to use linear regression. That being said, the present results add to the literature by showing that the presence of episodic memory difficulties is significantly linked to poorer comprehension of ironic and sincere assertions as well as poorer second-order mentalizing capacities. Regarding executive functions, results are in line with previous data having demonstrated a relationship between these capacities and irony comprehension [
One interesting finding of the present study is the observation of a relationship between measures of episodic memory and the performance on both the SSICT and the Combined Stories Test. These results are consistent with the proposition of Moreau et al. [
One surprising result is that, in both MCI and HC groups, ironic assertions were better interpreted than sincere ones. An analysis of the error patterns indicated that sincere stories were often mistaken as being ironic, and vice versa. In MCI individuals, “ironic” responses consisted of 60% of the total number of errors for sincere scenarios, and this pattern was reversed for ironic scenarios (60% of the time, participants responded “sincere” instead of “ironic”). In the HC group, more than 68% of errors in sincere scenarios consisted of “ironic” responses. Again, the reverse pattern was found in ironic stories, these being identified as “sincere” 64% of the time. One explanation could rely on the fact that stories were read out loud by the participants and were not presented through a prerecorded and controlled voice. Thus, this administration procedure precluded the use of prosody to correctly differentiate sincere from ironic assertions. Prosody is known to be of considerable help when having to differentiate such statements [
Prior to this study, the adapted SSICT was validated in 75 young adults (mean age = 25.92; SD = 2.8) and results indicated that participants better understood sincere compared to ironic scenarios. The discrepancy between findings of the present study and those of the validation study suggests that the differentiation between irony and sincerity gets harder with aging. If so, this interpretation would be consistent with the fact that age impacts second-order mentalizing as measured by short stories, as well as text comprehension
The SSICT has strengths and weaknesses. On one hand, the use of mentalizing as a variable equally distributed in the sincere and ironic stories represents an important improvement in comparison with previous tasks. On the other hand, the standardization of mentalizing between ironic and sincere scenarios may have explained to some extent the overall deficit of MCI participants in the present study. In MCI, verbal irony comprehension could probably be better assessed by means of tasks using different modalities (e.g., videotaped scenarios or other ecologically oriented tasks). This question requires further investigation.
A limitation of the present study relates to the generalization of the results to all MCI cases, considering the rather small sample size. Because of the known heterogeneity of MCI, the sample in this study may not reflect all the characteristics that can be found in this population. Thus, mentalizing and verbal irony comprehension in MCI should be studied using larger samples. Moreover, longitudinal follow-up is warranted to determine to what extent these cognitive functions can help predict the dementia stage of AD in those individuals having MCI. In all, it is possible that the level of education influenced the performance of our participants, as a link between education and mentalizing was already demonstrated in the aging population [
Overall, this study provides additional evidence of second-order mentalizing impairment in older adults with MCI and suggests that mental state inferences deficits are associated with difficulties in the comprehension of ironic and sincere assertions in this population. It also brings support to the possible use of mentalization tests in the screening of individuals at risks of evolving towards AD, although this requires further investigation. Further studies should aim at improving the comprehension of the cognitive correlates of ironic and sincere assertions comprehension (e.g., executive functions and episodic memory) and their implication in second-order mentalizing capacities of individuals with MCI or healthy older adults. Such work will not only help identifying early the prodromal stage of AD but also help clarifying the theoretical grounds of the comprehension of verbal irony and sincerity.
The authors declare that there is no conflict of interests regarding the publication and the results of this paper.
This study was funded by a research grant from the Alzheimer Society of Canada (#10 59) awarded to C. Hudon. G. Gaudreau received a doctoral scholarship from the Alzheimer Society of Canada (#11 58) and C. Hudon receives a salary award (