bs_bs_banner
PSYCHOGERIATRICS 2014; 14: 222–228
doi:10.1111/psyg.12101
ORIGINAL ARTICLE
Non-literal language deficits in mild cognitive impairment Sandra CARDOSO,1,2 Dina SILVA,2 João MAROCO,3 Alexandre DE MENDONÇA2 and Manuela GUERREIRO2
1
Institute of Health Sciences, Catholic University of Portugal, 2Faculty of Medicine and Institute of Molecular Medicine, University of Lisbon, and 3 Superior Institute of Applied Psychology, Lisbon, Portugal Correspondence: Prof Alexandre de Mendonça MD, PhD, Laboratory of Neurosciences, Institute of Molecular Medicine, Av. Prof. Egas Moniz, 1649-028 Lisbon, Portugal. Email: [email protected] Received 12 November 2013; revision received 20 May 2014; accepted 3 June 2014.
Key words: idiomatic expressions, literal language, mild cognitive impairment, non-literal language, proverbs, text comprehension.
Abstract Background: Verbal language deteriorates in Alzheimer’s disease, contributing to dramatic disturbances in the ability to communicate. The presence of language disturbances may be detected at earlier phases of the neurodegenerative process, like mild cognitive impairment (MCI). In daily verbal interactions, people mostly use literal language, but sometimes they employ non-literal language, which requires listeners to interpret messages beyond the plain meaning of the words and can be quite demanding. In the present study, we tested the hypotheses that patients with MCI may have deficits in non-literal language, and these deficits are more pronounced than deficits in literal language. Methods: Participants were recruited in a private memory clinic and senior universities. General cognitive evaluation included a comprehensive neuropsychological battery, the Mini-Mental State Examination, and the instrumental activities of daily living scale. Literal language was assessed with the semantic decision test, Token Test, and literal text comprehension test, and non-literal language with the proverbs comprehension, idiomatic expressions and non-literal text comprehension tests. Results: Fifty-two participants with MCI and 31 controls were recruited. Patients with MCI had lower scores in all complex language tests, both literal (Token Test, semantic decision and literal text) and non-literal (proverbs, idiomatic expressions, and non-literal text), than the controls; the difference in literal text score was marginally significant. As much as 69% of MCI participants had deficits (performance below 1.5 SD of the mean) on at least one of the complex language tasks. Deficits were more frequent on the proverbs comprehension and semantic decision tests, and the deficits on these tests did not significantly differ from that on the Token Test. Conclusion: Patients with MCI are hindered in understanding complex language, both literal and non-literal. In daily living, these complex language deficits could compromise effective verbal interactions with the others. Amelioration of these deficits should be an important intervention target as part of a comprehensive rehabilitation strategy for patients with cognitive decline.
INTRODUCTION Verbal language is one of the cognitive capacities that deteriorate in Alzheimer’s disease (AD), contributing to the dramatic disturbance in the ability to communicate often observed in patients with this disorder.1,2 Early in the disease process, speech is usually fluent, and the capacity to comprehend what is read and 222
heard is relatively maintained; however, later on, written and verbal comprehension are affected and oral production becomes less fluent, and at advanced stages, the patient may develop mutism.2–4 Clearly, the presence of language disturbances may be detected in the initial phases of the neurodegenerative process, like mild cognitive © 2014 The Authors Psychogeriatrics © 2014 Japanese Psychogeriatric Society
Non-literal language deficits in MCI
impairment (MCI), a condition likely to correspond to a prodromal stage of AD.5,6 The language deficits found in MCI patients may be qualitatively similar to those existing in AD patients, albeit more subtle.7 Several studies have showed that patients with MCI have deficits in tasks of verbal fluency and confrontation naming,7 as well as alterations in tests of language comprehension, such as the Token Test.8,9 In daily verbal interactions, people mostly use a type of language that is literal and, thus, more precise, direct, and easily understood. In contrast, non-literal language is mainly used in a ‘familiar’ context when the interlocutor wants to convey a message indirectly or disguise it so that the interpretation is not easily accessed. Non-literal language is a form of linguistic speech that demands the ability to interpret beyond the literal meaning of the message.10–12 It comprises the use of proverbs, hyperboles, metaphors, idiomatic expressions, irony, sarcasm, metonym, similes, personifications, hyperboles, onomatopoeias, and symbolism.10,12–15 It could be anticipated that patients with cognitive decline might have more prominent difficulties in dealing with non-literal language than with literal language. Patients with AD have deficits in identifying the correct meaning of a proverb when presented with multiple choices.16 Similarly, when confronted with images that include both the literal and non-literal meaning of an idiomatic expression, they identify the more straightforward literal meaning.17 Patients with AD also have difficulties in correctly explaining the intention and meaning of proverbs and idiomatic expressions.18–21 Few studies have analyzed non-literal language in patients at early stages of AD. Patients with amnestic MCI, when confronted with the need to explain proverbs, answered with more concrete but nonsensical answers than controls.22,23 However, it is not clear whether they also have difficulties in other non-literal language tasks, such as idiomatic expressions. Previous studies have not directly compared performances on literal and non-literal language tasks. In the present study, we tested two hypotheses: (i) patients with MCI have deficits in non-literal language tasks (proverbs, idiomatic expressions, and non-literal text comprehension); and (ii) patients with MCI have more difficulties with non-literal tasks than with literal language tasks. © 2014 The Authors Psychogeriatrics © 2014 Japanese Psychogeriatric Society
METHODS Research participants The participants were recruited in a private memory clinic (MCI participants) and senior universities (healthy control participants) in Lisbon. Inclusion criteria for both groups • Native Portuguese speakers • Education ≥ 4 years • Age 50–85 years old Inclusion criteria for the MCI group These criteria were adapted from the MCI criteria of the European Consortium on Alzheimer’s Disease.6 For reasons of sample homogeneity, only patients with amnestic MCI were recruited. The criteria were as follows: 1 Cognitive complaints came from the patients or their families. 2 The patient or informant reported a decline in cognitive functioning relative to previous abilities during the past year. 3 The patient had objective memory impairment as determined by the score on the immediate free recall of ‘Story A’ from the Logical Memory subtest of the Wechsler Memory Scale.24 Impairment was indicated by a score that was less than 1.5 SD the normative value for age and education. 4 The patient generally maintained activities of daily living, could participate in professional, social, and familial activities according to clinical evaluations, and had no or only mild impairment according to the instrumental activities of daily living (IADL) scale (i.e. no more than one item from the IADL scale was abnormal).25–27 Inclusion criteria for the control group • Absence of cognitive complaints • Normal score on immediate free recall of ‘Story A’ from the Logical Memory subtest of Wechsler Memory Scale24 • No impairment according to the IADL scale (i.e. no item from the IADL scale was abnormal)25–27 Exclusion criteria for both groups • Presence of neurological (stroke, brain tumour, significant head trauma, epilepsy), psychiatric disorders that may induce cognitive deficits, or major 223
S. Cardoso et al.
depression or serious depressive symptoms, as indicated by a score >10 points on the 15-item Geriatric Depression Scale28,29 • Systemic illness with cerebral impact (metabolic, endocrine, toxic, or infectious diseases) • History of alcohol and/or drug abuse • Presence of dementia or abnormal performance in the Mini-Mental State Examination, as indicated by a score below 22 for subjects with ≤11 years of education and below 27 for subjects with >11 years of education 30,31 Procedures All neuropsychological assessments were carried out by the same team of researchers (neuropsychologist and speech language therapist) in a quiet room for a period no longer than 45 min. Participants with MCI underwent a standard protocol comprising clinical history, neurological examination, laboratorial evaluation and brain imaging (computed tomography scan or nuclear magnetic resonance imaging).32 The study was conducted in accordance with the Declaration of Helsinki and approved by the local ethics committee. Participants completed the Portuguese versions of the following instruments: 1 Mini-Mental State Examination. This is widely used for a brief evaluation of mental state and to screen for dementia.30 The normative cut-off values for the Portuguese population adjusted to education were used.31 2 instrumental activities of daily living scale (IADL).25,26 IADL score reflects the number of impaired or maintained activities of daily living. 3 Bateria de Lisboa para Avaliação das Demências. Neuropsychological assessment was done through this test,33,34 a comprehensive neuropsychological battery evaluating multiple cognitive domains that has been validated for the Portuguese population. This battery includes tests for the following cognitive domains: attention (cancellation task); verbal initiative (semantic fluency); motor and graphomotor initiatives; verbal comprehension (a modified version of the Token Test); verbal and nonverbal abstraction (interpretation of proverbs and the Raven’s Colour Matrices – Ab series-B); orientation (personal, spatial, and temporal); visuoconstructional abilities (cube copy); planning and visuospatial/praxis abilities (clock draw); calculation (basic written calculation); object naming; object 224
4
5
6
7
8
identification; attention and working memory (digit span forward, digit span backward); visual memory (visual reproduction test); learning and verbal memory (verbal paired-associate learning, word recall); and episodic memory (Logical Memory immediate and delayed). Participants with MCI underwent the complete battery, but only episodic and working memory tests (Logical Memory immediate and delayed, digit span forward and backward) were applied to the controls. semantic decision, a test from a battery of assessments of complex language (Provas de Avaliação de Linguagem Complexa (PLINC)).35 In the semantic decision task, the participant must identify the relationship between 30 word pairs, namely identical meaning, opposite meaning, or no relation at all. The maximum score is 30, 1 point for each item.36 Token Test (22-item short version).37–40 This test evaluates comprehension, with orders of increasing length and complexity. The maximum score is 22, 1 for each order. literal text comprehension test, from a comprehensive aphasia battery (Bateria de Avaliação da Afasia de Lisboa).38–40 This test evaluates participants’ comprehension of written material; participants read a passage silently and then answer six questions. The text remains in front of participants while they answer the questions. The maximum score is 6, 1 point for each question. proverbs comprehension, a test from PLINC.35 Participants are presented with six sets of proverbs during this test. For each set, the participant is presented with a target proverb and must match it to a proverb with a similar meaning from a choice of three: a non-related proverb, a proverb with the same figurative meaning (the correct answer), and a proverb containing similar semantic concepts but with a different meaning. The participant is asked to choose the proverb with the closest meaning to the target proverb. The maximum score is 6, 1 point for each set of proverbs.36 idiomatic expressions, a test from PLINC.35 This test evaluates a participant’s capacity to explain an abstract concept, and it includes six items. Comprehension is evaluated according to the degree of abstraction: 3 points for answers with the maximum level of abstraction, 2 points for the intermediate level, and 1 point for answers with a literal meaning, and 0 points for answers not related to the item or © 2014 The Authors Psychogeriatrics © 2014 Japanese Psychogeriatric Society
Non-literal language deficits in MCI
failure to answer. The score ranges between 0 and 18.36 9 non-literal text comprehension test, from the Protocole Montréal d’Évaluation de la Communication.41,42 The purpose of this test is to evaluate the comprehension of non-literal language. Participants silently read a passage and then answer 12 questions. The text remains in front of participants while they answer the questions. The maximum score is 12, 1 point for each question. Statistical analysis The statistical analysis was performed using SPSS v.19 for Windows (SPSS, Chicago, IL, USA). P < 0.05 was considered statistically significant. The sample size of 52 patients with MCI and 31 controls (ratio 2:1) was calculated to detect a difference of 2.5 points in the idiomatic expressions test and a difference of 0.8 points in the proverbs test between MCI patients and controls (based on preliminary data from a pilot study with healthy, aged subjects), with a significance level of 0.05 and power of 0.90 (Power and Precision, version 4.1, Englewood, NJ, USA). Comparison of demographic and neuropsychological data between the two groups was performed using Student’s t test for numerical data and χ2 test for categorical data. Frequencies of deficits in literal and non-literal language tasks in MCI patients were compared with χ2 test and then Z-tests for differences between proportions as implemented in SPSS v.19 Crosstabs. Impairment on any test was defined as a score more than 1.5 SD below the mean for the age and education. Categorical principal components analysis was used to analyze the relation between language and other cognitive tasks. The criterion of retention for the components was according the eigenvalue rule (>1), and the internal consistence of each component was measured by the Cronbach’s α test.
RESULTS All subjects (n = 85) agreed to participate in the study. Two control participants were eventually excluded, one due to poor performance on the neuropsychological evaluation (Logical Memory and digit span values were below 1.5 SD of the normative data) and the other due to marked depressive symptoms. © 2014 The Authors Psychogeriatrics © 2014 Japanese Psychogeriatric Society
There were no significant differences in age, education, and gender between the two groups (Table 1). The MCI participants performed worse than the controls on the Mini-Mental State Examination and Logical Memory test (immediate and delayed recall) (Student’s t-test) (Table 1). In basic language capacities and naming and identifying objects, none of the participants showed deficits. MCI participants had worse scores on all complex language performance tests, both literal (Token Test, semantic decision, and literal text) and non-literal (proverbs, idiomatic expressions and non-literal text), than the controls (Student’s t-test) (Table 2); the difference in literal text score was marginally significant. The frequencies of deficits (performance below 1.5 SD of the mean) for the different language tasks were compared in MCI participants. As much as 69% of MCI participants had deficits on at least one of the complex language tasks. Deficits were more frequent on the semantic decision and proverbs comprehension tests, and the deficits on these tests did not significantly differ from that on the Token Test (χ2 test) (Table 3). Next, using categorical principal components statistical analysis, we investigated MCI patients’ performance on language tests and other selected cognitive domains. All language tasks were loaded to the first component, which we called ‘language’; the second component included episodic memory tasks (Logical Memory immediate recall and delayed recall); and the third component included attentional and working memory tasks (digit span forward and backward). This model could explain 66% of the total variance (Table 4).
Table 1 Demographic and neuropsychological characterization
Age (years) Education (years) Gender (M/F) (n) Mini-Mental State Examination Logical Memory immediate recall Logical Memory delayed recall Digit span forward Digit span backward
MCI (n = 52) mean 1 SD
Control (n = 31) mean 1 SD
P-value
71.8 1 8.2 10.4 1 3.6 22/30 26.1 1 2.6
70.1 1 7.4 11.4 1 2.4 7/24 29.1 1 0.8
0.103‡ 0.089‡ 0.096†
PSYCHOGERIATRICS 2014; 14: 222–228
doi:10.1111/psyg.12101
ORIGINAL ARTICLE
Non-literal language deficits in mild cognitive impairment Sandra CARDOSO,1,2 Dina SILVA,2 João MAROCO,3 Alexandre DE MENDONÇA2 and Manuela GUERREIRO2
1
Institute of Health Sciences, Catholic University of Portugal, 2Faculty of Medicine and Institute of Molecular Medicine, University of Lisbon, and 3 Superior Institute of Applied Psychology, Lisbon, Portugal Correspondence: Prof Alexandre de Mendonça MD, PhD, Laboratory of Neurosciences, Institute of Molecular Medicine, Av. Prof. Egas Moniz, 1649-028 Lisbon, Portugal. Email: [email protected] Received 12 November 2013; revision received 20 May 2014; accepted 3 June 2014.
Key words: idiomatic expressions, literal language, mild cognitive impairment, non-literal language, proverbs, text comprehension.
Abstract Background: Verbal language deteriorates in Alzheimer’s disease, contributing to dramatic disturbances in the ability to communicate. The presence of language disturbances may be detected at earlier phases of the neurodegenerative process, like mild cognitive impairment (MCI). In daily verbal interactions, people mostly use literal language, but sometimes they employ non-literal language, which requires listeners to interpret messages beyond the plain meaning of the words and can be quite demanding. In the present study, we tested the hypotheses that patients with MCI may have deficits in non-literal language, and these deficits are more pronounced than deficits in literal language. Methods: Participants were recruited in a private memory clinic and senior universities. General cognitive evaluation included a comprehensive neuropsychological battery, the Mini-Mental State Examination, and the instrumental activities of daily living scale. Literal language was assessed with the semantic decision test, Token Test, and literal text comprehension test, and non-literal language with the proverbs comprehension, idiomatic expressions and non-literal text comprehension tests. Results: Fifty-two participants with MCI and 31 controls were recruited. Patients with MCI had lower scores in all complex language tests, both literal (Token Test, semantic decision and literal text) and non-literal (proverbs, idiomatic expressions, and non-literal text), than the controls; the difference in literal text score was marginally significant. As much as 69% of MCI participants had deficits (performance below 1.5 SD of the mean) on at least one of the complex language tasks. Deficits were more frequent on the proverbs comprehension and semantic decision tests, and the deficits on these tests did not significantly differ from that on the Token Test. Conclusion: Patients with MCI are hindered in understanding complex language, both literal and non-literal. In daily living, these complex language deficits could compromise effective verbal interactions with the others. Amelioration of these deficits should be an important intervention target as part of a comprehensive rehabilitation strategy for patients with cognitive decline.
INTRODUCTION Verbal language is one of the cognitive capacities that deteriorate in Alzheimer’s disease (AD), contributing to the dramatic disturbance in the ability to communicate often observed in patients with this disorder.1,2 Early in the disease process, speech is usually fluent, and the capacity to comprehend what is read and 222
heard is relatively maintained; however, later on, written and verbal comprehension are affected and oral production becomes less fluent, and at advanced stages, the patient may develop mutism.2–4 Clearly, the presence of language disturbances may be detected in the initial phases of the neurodegenerative process, like mild cognitive © 2014 The Authors Psychogeriatrics © 2014 Japanese Psychogeriatric Society
Non-literal language deficits in MCI
impairment (MCI), a condition likely to correspond to a prodromal stage of AD.5,6 The language deficits found in MCI patients may be qualitatively similar to those existing in AD patients, albeit more subtle.7 Several studies have showed that patients with MCI have deficits in tasks of verbal fluency and confrontation naming,7 as well as alterations in tests of language comprehension, such as the Token Test.8,9 In daily verbal interactions, people mostly use a type of language that is literal and, thus, more precise, direct, and easily understood. In contrast, non-literal language is mainly used in a ‘familiar’ context when the interlocutor wants to convey a message indirectly or disguise it so that the interpretation is not easily accessed. Non-literal language is a form of linguistic speech that demands the ability to interpret beyond the literal meaning of the message.10–12 It comprises the use of proverbs, hyperboles, metaphors, idiomatic expressions, irony, sarcasm, metonym, similes, personifications, hyperboles, onomatopoeias, and symbolism.10,12–15 It could be anticipated that patients with cognitive decline might have more prominent difficulties in dealing with non-literal language than with literal language. Patients with AD have deficits in identifying the correct meaning of a proverb when presented with multiple choices.16 Similarly, when confronted with images that include both the literal and non-literal meaning of an idiomatic expression, they identify the more straightforward literal meaning.17 Patients with AD also have difficulties in correctly explaining the intention and meaning of proverbs and idiomatic expressions.18–21 Few studies have analyzed non-literal language in patients at early stages of AD. Patients with amnestic MCI, when confronted with the need to explain proverbs, answered with more concrete but nonsensical answers than controls.22,23 However, it is not clear whether they also have difficulties in other non-literal language tasks, such as idiomatic expressions. Previous studies have not directly compared performances on literal and non-literal language tasks. In the present study, we tested two hypotheses: (i) patients with MCI have deficits in non-literal language tasks (proverbs, idiomatic expressions, and non-literal text comprehension); and (ii) patients with MCI have more difficulties with non-literal tasks than with literal language tasks. © 2014 The Authors Psychogeriatrics © 2014 Japanese Psychogeriatric Society
METHODS Research participants The participants were recruited in a private memory clinic (MCI participants) and senior universities (healthy control participants) in Lisbon. Inclusion criteria for both groups • Native Portuguese speakers • Education ≥ 4 years • Age 50–85 years old Inclusion criteria for the MCI group These criteria were adapted from the MCI criteria of the European Consortium on Alzheimer’s Disease.6 For reasons of sample homogeneity, only patients with amnestic MCI were recruited. The criteria were as follows: 1 Cognitive complaints came from the patients or their families. 2 The patient or informant reported a decline in cognitive functioning relative to previous abilities during the past year. 3 The patient had objective memory impairment as determined by the score on the immediate free recall of ‘Story A’ from the Logical Memory subtest of the Wechsler Memory Scale.24 Impairment was indicated by a score that was less than 1.5 SD the normative value for age and education. 4 The patient generally maintained activities of daily living, could participate in professional, social, and familial activities according to clinical evaluations, and had no or only mild impairment according to the instrumental activities of daily living (IADL) scale (i.e. no more than one item from the IADL scale was abnormal).25–27 Inclusion criteria for the control group • Absence of cognitive complaints • Normal score on immediate free recall of ‘Story A’ from the Logical Memory subtest of Wechsler Memory Scale24 • No impairment according to the IADL scale (i.e. no item from the IADL scale was abnormal)25–27 Exclusion criteria for both groups • Presence of neurological (stroke, brain tumour, significant head trauma, epilepsy), psychiatric disorders that may induce cognitive deficits, or major 223
S. Cardoso et al.
depression or serious depressive symptoms, as indicated by a score >10 points on the 15-item Geriatric Depression Scale28,29 • Systemic illness with cerebral impact (metabolic, endocrine, toxic, or infectious diseases) • History of alcohol and/or drug abuse • Presence of dementia or abnormal performance in the Mini-Mental State Examination, as indicated by a score below 22 for subjects with ≤11 years of education and below 27 for subjects with >11 years of education 30,31 Procedures All neuropsychological assessments were carried out by the same team of researchers (neuropsychologist and speech language therapist) in a quiet room for a period no longer than 45 min. Participants with MCI underwent a standard protocol comprising clinical history, neurological examination, laboratorial evaluation and brain imaging (computed tomography scan or nuclear magnetic resonance imaging).32 The study was conducted in accordance with the Declaration of Helsinki and approved by the local ethics committee. Participants completed the Portuguese versions of the following instruments: 1 Mini-Mental State Examination. This is widely used for a brief evaluation of mental state and to screen for dementia.30 The normative cut-off values for the Portuguese population adjusted to education were used.31 2 instrumental activities of daily living scale (IADL).25,26 IADL score reflects the number of impaired or maintained activities of daily living. 3 Bateria de Lisboa para Avaliação das Demências. Neuropsychological assessment was done through this test,33,34 a comprehensive neuropsychological battery evaluating multiple cognitive domains that has been validated for the Portuguese population. This battery includes tests for the following cognitive domains: attention (cancellation task); verbal initiative (semantic fluency); motor and graphomotor initiatives; verbal comprehension (a modified version of the Token Test); verbal and nonverbal abstraction (interpretation of proverbs and the Raven’s Colour Matrices – Ab series-B); orientation (personal, spatial, and temporal); visuoconstructional abilities (cube copy); planning and visuospatial/praxis abilities (clock draw); calculation (basic written calculation); object naming; object 224
4
5
6
7
8
identification; attention and working memory (digit span forward, digit span backward); visual memory (visual reproduction test); learning and verbal memory (verbal paired-associate learning, word recall); and episodic memory (Logical Memory immediate and delayed). Participants with MCI underwent the complete battery, but only episodic and working memory tests (Logical Memory immediate and delayed, digit span forward and backward) were applied to the controls. semantic decision, a test from a battery of assessments of complex language (Provas de Avaliação de Linguagem Complexa (PLINC)).35 In the semantic decision task, the participant must identify the relationship between 30 word pairs, namely identical meaning, opposite meaning, or no relation at all. The maximum score is 30, 1 point for each item.36 Token Test (22-item short version).37–40 This test evaluates comprehension, with orders of increasing length and complexity. The maximum score is 22, 1 for each order. literal text comprehension test, from a comprehensive aphasia battery (Bateria de Avaliação da Afasia de Lisboa).38–40 This test evaluates participants’ comprehension of written material; participants read a passage silently and then answer six questions. The text remains in front of participants while they answer the questions. The maximum score is 6, 1 point for each question. proverbs comprehension, a test from PLINC.35 Participants are presented with six sets of proverbs during this test. For each set, the participant is presented with a target proverb and must match it to a proverb with a similar meaning from a choice of three: a non-related proverb, a proverb with the same figurative meaning (the correct answer), and a proverb containing similar semantic concepts but with a different meaning. The participant is asked to choose the proverb with the closest meaning to the target proverb. The maximum score is 6, 1 point for each set of proverbs.36 idiomatic expressions, a test from PLINC.35 This test evaluates a participant’s capacity to explain an abstract concept, and it includes six items. Comprehension is evaluated according to the degree of abstraction: 3 points for answers with the maximum level of abstraction, 2 points for the intermediate level, and 1 point for answers with a literal meaning, and 0 points for answers not related to the item or © 2014 The Authors Psychogeriatrics © 2014 Japanese Psychogeriatric Society
Non-literal language deficits in MCI
failure to answer. The score ranges between 0 and 18.36 9 non-literal text comprehension test, from the Protocole Montréal d’Évaluation de la Communication.41,42 The purpose of this test is to evaluate the comprehension of non-literal language. Participants silently read a passage and then answer 12 questions. The text remains in front of participants while they answer the questions. The maximum score is 12, 1 point for each question. Statistical analysis The statistical analysis was performed using SPSS v.19 for Windows (SPSS, Chicago, IL, USA). P < 0.05 was considered statistically significant. The sample size of 52 patients with MCI and 31 controls (ratio 2:1) was calculated to detect a difference of 2.5 points in the idiomatic expressions test and a difference of 0.8 points in the proverbs test between MCI patients and controls (based on preliminary data from a pilot study with healthy, aged subjects), with a significance level of 0.05 and power of 0.90 (Power and Precision, version 4.1, Englewood, NJ, USA). Comparison of demographic and neuropsychological data between the two groups was performed using Student’s t test for numerical data and χ2 test for categorical data. Frequencies of deficits in literal and non-literal language tasks in MCI patients were compared with χ2 test and then Z-tests for differences between proportions as implemented in SPSS v.19 Crosstabs. Impairment on any test was defined as a score more than 1.5 SD below the mean for the age and education. Categorical principal components analysis was used to analyze the relation between language and other cognitive tasks. The criterion of retention for the components was according the eigenvalue rule (>1), and the internal consistence of each component was measured by the Cronbach’s α test.
RESULTS All subjects (n = 85) agreed to participate in the study. Two control participants were eventually excluded, one due to poor performance on the neuropsychological evaluation (Logical Memory and digit span values were below 1.5 SD of the normative data) and the other due to marked depressive symptoms. © 2014 The Authors Psychogeriatrics © 2014 Japanese Psychogeriatric Society
There were no significant differences in age, education, and gender between the two groups (Table 1). The MCI participants performed worse than the controls on the Mini-Mental State Examination and Logical Memory test (immediate and delayed recall) (Student’s t-test) (Table 1). In basic language capacities and naming and identifying objects, none of the participants showed deficits. MCI participants had worse scores on all complex language performance tests, both literal (Token Test, semantic decision, and literal text) and non-literal (proverbs, idiomatic expressions and non-literal text), than the controls (Student’s t-test) (Table 2); the difference in literal text score was marginally significant. The frequencies of deficits (performance below 1.5 SD of the mean) for the different language tasks were compared in MCI participants. As much as 69% of MCI participants had deficits on at least one of the complex language tasks. Deficits were more frequent on the semantic decision and proverbs comprehension tests, and the deficits on these tests did not significantly differ from that on the Token Test (χ2 test) (Table 3). Next, using categorical principal components statistical analysis, we investigated MCI patients’ performance on language tests and other selected cognitive domains. All language tasks were loaded to the first component, which we called ‘language’; the second component included episodic memory tasks (Logical Memory immediate recall and delayed recall); and the third component included attentional and working memory tasks (digit span forward and backward). This model could explain 66% of the total variance (Table 4).
Table 1 Demographic and neuropsychological characterization
Age (years) Education (years) Gender (M/F) (n) Mini-Mental State Examination Logical Memory immediate recall Logical Memory delayed recall Digit span forward Digit span backward
MCI (n = 52) mean 1 SD
Control (n = 31) mean 1 SD
P-value
71.8 1 8.2 10.4 1 3.6 22/30 26.1 1 2.6
70.1 1 7.4 11.4 1 2.4 7/24 29.1 1 0.8
0.103‡ 0.089‡ 0.096†
