**2. Material and methods**

#### **2.1. Material**

The battery of Neuropsychological Tests Abbreviated and Adapted for Spanish Speakers, a valid and reliable instrument developed to detect dementia, aging and cognitive impairment, including the probable site of brain impairment, was administered [52–55]. Sixtyseven indicators of 25 basic subtests were analyzed. The present battery assessed the task completion time [i.e., the processing time (T)] in several subtests as well as: (1) spontaneous speech (in its aphasic manifestations); (2) personal orientation; (3) time and place orientation, and errors (E) in time orientation; (4) phonemic discrimination (letter 'A') by auditory cancelation (verbal auditory selective and sustained attention: omission and commission E); (5) figure discrimination (triangle) by visual cancelation [nonverbal visual selective and sustained attention: correct responses (CR) as well as errors and time (E&T)]; (6) direct and reverse serial order (months forward and backwards: E&T); (7) spatial memory (five hidden objects: accuracy (remembered objects and places) as well as four different types of E); (8) copy of alternating or repetitive graph series; (9) copy and naming (written response) of simple figures; (10) constructional praxia (cube and clock drawing in response to commands: CR and T); (11) syntax-complex verbal comprehension; (12) verbal auditory attention span (digits: forward and backwards); (13) writing abilities such as writing one verbal automatism (the name), writing by copying and by dictation and writing-legibility; (14) written verbal fluency [quantity: number of words, quality: syntactic complexity, legibility: overall score and legibility regardless of quantity (average score per word)]; (15) written arithmetic operations; (16) mental calculations (subtracting serial sevens: CR and T); (17) oral verbal fluency (number of words beginning with 'F'); (18) reading (a story): oral expression and abstraction/comprehension; (19) visual memory: face recognition; (20) visual memory: retrieval of a complex figure; (21) graphesthesia; (22) finger recognition; (23) a delayed story recall (spontaneous and cued, using two indicators: the interviewer's global impression during administration, and a standardized and detailed scoring of 25 passages after administration); (24) the paired-associate word learning, which included three trials and a delayed recall of easy and hard pairs and (25) semantic verbal memory/naming by picture confrontation. [Note: In general, accuracy (CR) was assessed unless otherwise indicated by E, T, and/or E&T.] The tasks of the battery related with DP were complex verbal comprehension (i.e., syntax-complex verbal comprehension and story comprehension) in addition to storytelling [i.e., a delayed story recall (spontaneous and cued)]. The tasks of the battery related with fluency were written verbal fluency (quantity: number of words) and oral verbal fluency (number of words beginning with 'F'). Details of test administration and scores are explained elsewhere [49–56].

The present study is part of a bigger research project which aims at developing efficient tests, that is, brief and/or easy to apply neuropsychological techniques without neglecting the goals of accuracy and validity (see, e.g., [49–51]). Since theory and validity are interlaced, it is expected that the present data are not only useful to hypothesize about the bases of TS, but also to explore the viability, validity and reliability of the present scale to assess TS in a natu-

In summary, the present study aimed to explore, in patients with focal brain injuries, if TS is associated with cognitive, emotional or behavioral impairments and with specific sites of brain injury. Complementarily, the present study aimed to explore if a hypothetical pattern of neuropsychological and/or neuroanatomical impairments can be identified for TS as well as if

In view of the reviewed, and in an attempt to delimit the conceptual definition of TS, only communication dysfunctions which affect the quality and consistency of information in the topic of the discourse, without affecting the most basic resources to carry out such discourse, were considered. More specifically, when: (a) the deficit was secondary to brain injury, (b) the patient was alert, without aphasia, without psychiatric history and without TS history and (c) according to the conditions which were expressed in the first paragraph of this work, the topic of the conversation was missing (i.e., the topic was irrelevant to the interview situation, or it was not well preserved or focused during the interaction) the resulting speech was defined

In view of the exploratory nature of the study, a comprehensive neuropsychological battery was administered because all the battery tests and subtests were in principle considered potential factors for explaining TS. However, and bearing in mind that the tasks of narrative comprehension, memory and production have been previously recognized as valid measures of discourse processing (DP) [8–10], they were specially evaluated. Considering that tests of fluency have been used as indicators of logorrhoea [6], and that logorrhoea includes failures in the quality and consistency of speech, the performance in tasks of spoken and written ver-

The battery of Neuropsychological Tests Abbreviated and Adapted for Spanish Speakers, a valid and reliable instrument developed to detect dementia, aging and cognitive impairment, including the probable site of brain impairment, was administered [52–55]. Sixtyseven indicators of 25 basic subtests were analyzed. The present battery assessed the task completion time [i.e., the processing time (T)] in several subtests as well as: (1) spontaneous speech (in its aphasic manifestations); (2) personal orientation; (3) time and place orientation, and errors (E) in time orientation; (4) phonemic discrimination (letter 'A') by auditory

reliability and validity indices can be obtained for the present TS screening scale.

ral situation, by the bedside of the patient.

as tangential.

196 Gerontology

**2.1. Material**

bal fluency were also evaluated.

**2. Material and methods**

The emergence of the following disorders as a consequence of brain injury as reported by the caregiver during the initial interview were also registered (scale range: 0–3): sensory deficits; motor deficits; perceptual-cognitive disorders (i.e., difficulty in recognizing known persons, places, moments or objects, independently of sensory acuity); sleeping disorders (i.e., insomnia, somnolence during the day, etc.); language disorders (i.e., paraphasias, anomies, echolalia, intrusions, reduced verbal comprehension or fluency, dysarthria, etc.); behavioral disorders (i.e., abnormal responses, anxiety, irritability, depression, lack of sphincter control, difficulty in organizing action, changes of personality, etc.); and thought disturbances (i.e., hallucinations, delusions, loss of sense of reality, dissociative symptoms, etc.). The presence of seizures was also registered.

Some complementary behavioral observations, which are usually evaluated during the administration of the comprehensive battery were also analyzed: the behavioral observations computed in this study were: degree of cooperation (0–3, i.e., absent: 0, very poor: 1, poor: 2, good: 3); emotional state (−1 to 1, i.e., inhibited: −1, normal: 0, excited: 1); disability awareness (0–3, i.e., null: 0, bad: 1, regular: 2, good: 3); language speed (−1 to 1, i.e., slow: −1, normal: 0, rapid: +1); voice volume (0–4, i.e., whispered: 0, hypophonic: 1, low: 2, normal: 3, hyperphonic: 4) and prosody (0–3, i.e., total or severe dysprosody: 0, moderate prosody: 1, slight prosody: 2, normal expression or prosody: 3). The presence of emotional lability, aggression, hallucinations, delusions and verbal perseverations (including words and/or thoughts) was also registered.

The emergence of TS was registered as a feature of spontaneous speech, different from the aphasia symptoms usually assessed by this item. In the item of spontaneous speech, the patients' ability to describe their own disease is explored. The interviewer's question in the TS item was: *'Tell me what happened to you and why you are here. (When did the problem start? How* 

Neurocognitive Implications of Tangential Speech in Patients with Focal Brain Damage

http://dx.doi.org/10.5772/intechopen.71904

199

If the topic of the conversation was missing at any moment of the interaction, and digressive responses were maintained irreversibly, without spontaneous recovering, throughout the three successive statements, the resulting speech was empirically defined as tangential. TS was coded as present. Subsequently, the interviewer gave a prompt. (In order to corroborate the interviewees' own ability to get back to the point, interviewers did not have to give indications or ask questions which facilitate the recall of the topic of the conversation.) If TS was

0 = Empty talk; pointless speech (the thread of the conversation is missing); inconsistency with the context and with a line of communication; disconnected from the listener; permanent

1 = Speech disconnected from the goal of the conversation, or difficult to insert into a coherent line of communication, most of the time or in most of the expressions; interviewee may or may not get back to the point by means of an interviewer's prompt such as 'and so?' The

2 = Speech that may drift into nonessential details without straying too far from the main topic of conversation. Although it has a fluctuating direction (sometimes it approaches the topic and sometimes it scatters for no apparent reason), the interviewee can usually get back to the point at the request of the interviewer. The discourse is rarely impaired. Topic recover-

3 = Correct or normal speech in its logical sense and adequacy to the context. If at times it deviates a bit from the topic, involving marginal comments, the main idea or gestalt returns

The interviewee's verbatim response was recorded (hand-written format) and the interviewer's prompts or questions were registered with a vertical line. Transcripts were reanalyzed by a second rater to assess inter-rater reliability. In order to carry out this study, transcripts were rated blindly by two trained neuropsychologists, members of the research team. [Note: As the prompts for TS cannot be changed, each TS prompt was binary-coded as 0 (disagreement) or 1 (agreement). The reason for that was to avoid magnifying the correlation by reevaluating only those patients within the range of TS. If all the indications were reassessed with a value of 0 (100% disagreement), the patient's score (initially <3) was increased by one point. If no prompt was provided in the first instance, and in the second evaluation, it was thought that the interviewee should have received at least one prompt, the patient's score (initially = 3) was

repeated three times (maximum four prompts), the interview was finished.

irrelevant comments. The discourse is impaired. Topic recovering is 0%.

discourse is relatively or mostly impaired. Topic recovering is >0% and ≤50%.

*was it…?).'*

ing is >50%.

decreased in one point].

The following scale was applied:

spontaneously. The discourse is not impaired.

#### **2.2. Subjects and procedures**

Data were obtained from a sample of 175 Argentine Spanish-speaking right-handed volunteers. Clinical data were obtained from a sample of 95 patients who were consecutively recruited from the Neurological and Neurosurgery Service of the Cordoba Hospital, a public hospital for adults. Demographically matched healthy participants (HP) were recruited from cultural, recreational and retirement centers in the province of Cordoba. HP were included if they were independent and adapted to daily life demands, without any known neurological or psychiatric disease. HP were excluded if they had: (i) TS or any type of language impairment, (ii) symptoms of neurological or psychiatric disorders, (iii) risk of neurological damage by disease or accident, (iv) any kind of medical condition which could affect neuropsychological performance or (v) sensorial or motor difficulties which could prevent them from carrying out the tests fluently. The recruitment method is better described elsewhere [49–51, 54, 57]. Patients were included if they had focal brain lesions confirmed by MRI and complementary diagnostic studies, and if they were preoperative inpatients. Patients were excluded if they: (i) had multiple or diffuse brain damage, (ii) had any other (previous or simultaneous) associated neurological disease, (iii) had history of psychiatric disorders, (iv) had history of TS, (v) were treated with psychotropic medication, (vi) had aphasia, hemianopia, hemineglect, hemihypesthesia or minimum signs of clouding of consciousness, according to the coincident report among the physician (before administering the battery), the caregiver (during the initial interview), as well as the neuropsychologist (during the administration of the battery). The data collected during the initial interview with the caregivers was taken as evidence of the premorbid condition. The comprehensive neuropsychological battery was administered and scored blindly to neuroanatomical data and the TS scale, which was applied by other member of the research team.

Patients grouped by TS were compared on their demographic variables as well as on type and site/side of lesion, disease duration (reported in months), risk factors (malnutrition, frequent contact with toxic agents, hypertension, heart disease, obesity, diabetes, genetic component of the illness, alcohol or drug consumption, etc.), and the presence of brachial and crural hemiparesis. Regarding the sites of lesion, they were divided into anterior hemisphere (frontal) lesions (A) versus posterior hemisphere (temporal, parietal or occipital) lesions (P). Lesions located in inferior structures (such as thalamus, basal ganglia, internal capsule, etc.) were classified as subcortical (SC) lesions; and lesions located in the frontal lobe and any of the posterior lobes, or in regions located between the frontal lobe and the posterior lobes, were classified as antero-posterior (AP) ones. As well, lesions were divided into left (L), right (R) and bilateral (B), according the injured hemisphere.

The emergence of TS was registered as a feature of spontaneous speech, different from the aphasia symptoms usually assessed by this item. In the item of spontaneous speech, the patients' ability to describe their own disease is explored. The interviewer's question in the TS item was: *'Tell me what happened to you and why you are here. (When did the problem start? How was it…?).'*

If the topic of the conversation was missing at any moment of the interaction, and digressive responses were maintained irreversibly, without spontaneous recovering, throughout the three successive statements, the resulting speech was empirically defined as tangential. TS was coded as present. Subsequently, the interviewer gave a prompt. (In order to corroborate the interviewees' own ability to get back to the point, interviewers did not have to give indications or ask questions which facilitate the recall of the topic of the conversation.) If TS was repeated three times (maximum four prompts), the interview was finished.

The following scale was applied:

i.e., null: 0, bad: 1, regular: 2, good: 3); language speed (−1 to 1, i.e., slow: −1, normal: 0, rapid: +1); voice volume (0–4, i.e., whispered: 0, hypophonic: 1, low: 2, normal: 3, hyperphonic: 4) and prosody (0–3, i.e., total or severe dysprosody: 0, moderate prosody: 1, slight prosody: 2, normal expression or prosody: 3). The presence of emotional lability, aggression, hallucinations, delusions and verbal perseverations (including words and/or thoughts) was also registered.

Data were obtained from a sample of 175 Argentine Spanish-speaking right-handed volunteers. Clinical data were obtained from a sample of 95 patients who were consecutively recruited from the Neurological and Neurosurgery Service of the Cordoba Hospital, a public hospital for adults. Demographically matched healthy participants (HP) were recruited from cultural, recreational and retirement centers in the province of Cordoba. HP were included if they were independent and adapted to daily life demands, without any known neurological or psychiatric disease. HP were excluded if they had: (i) TS or any type of language impairment, (ii) symptoms of neurological or psychiatric disorders, (iii) risk of neurological damage by disease or accident, (iv) any kind of medical condition which could affect neuropsychological performance or (v) sensorial or motor difficulties which could prevent them from carrying out the tests fluently. The recruitment method is better described elsewhere [49–51, 54, 57]. Patients were included if they had focal brain lesions confirmed by MRI and complementary diagnostic studies, and if they were preoperative inpatients. Patients were excluded if they: (i) had multiple or diffuse brain damage, (ii) had any other (previous or simultaneous) associated neurological disease, (iii) had history of psychiatric disorders, (iv) had history of TS, (v) were treated with psychotropic medication, (vi) had aphasia, hemianopia, hemineglect, hemihypesthesia or minimum signs of clouding of consciousness, according to the coincident report among the physician (before administering the battery), the caregiver (during the initial interview), as well as the neuropsychologist (during the administration of the battery). The data collected during the initial interview with the caregivers was taken as evidence of the premorbid condition. The comprehensive neuropsychological battery was administered and scored blindly to neuroanatomical data and the TS scale, which was applied by other

Patients grouped by TS were compared on their demographic variables as well as on type and site/side of lesion, disease duration (reported in months), risk factors (malnutrition, frequent contact with toxic agents, hypertension, heart disease, obesity, diabetes, genetic component of the illness, alcohol or drug consumption, etc.), and the presence of brachial and crural hemiparesis. Regarding the sites of lesion, they were divided into anterior hemisphere (frontal) lesions (A) versus posterior hemisphere (temporal, parietal or occipital) lesions (P). Lesions located in inferior structures (such as thalamus, basal ganglia, internal capsule, etc.) were classified as subcortical (SC) lesions; and lesions located in the frontal lobe and any of the posterior lobes, or in regions located between the frontal lobe and the posterior lobes, were classified as antero-posterior (AP) ones. As well, lesions were divided into left (L), right (R)

**2.2. Subjects and procedures**

198 Gerontology

member of the research team.

and bilateral (B), according the injured hemisphere.

0 = Empty talk; pointless speech (the thread of the conversation is missing); inconsistency with the context and with a line of communication; disconnected from the listener; permanent irrelevant comments. The discourse is impaired. Topic recovering is 0%.

1 = Speech disconnected from the goal of the conversation, or difficult to insert into a coherent line of communication, most of the time or in most of the expressions; interviewee may or may not get back to the point by means of an interviewer's prompt such as 'and so?' The discourse is relatively or mostly impaired. Topic recovering is >0% and ≤50%.

2 = Speech that may drift into nonessential details without straying too far from the main topic of conversation. Although it has a fluctuating direction (sometimes it approaches the topic and sometimes it scatters for no apparent reason), the interviewee can usually get back to the point at the request of the interviewer. The discourse is rarely impaired. Topic recovering is >50%.

3 = Correct or normal speech in its logical sense and adequacy to the context. If at times it deviates a bit from the topic, involving marginal comments, the main idea or gestalt returns spontaneously. The discourse is not impaired.

The interviewee's verbatim response was recorded (hand-written format) and the interviewer's prompts or questions were registered with a vertical line. Transcripts were reanalyzed by a second rater to assess inter-rater reliability. In order to carry out this study, transcripts were rated blindly by two trained neuropsychologists, members of the research team. [Note: As the prompts for TS cannot be changed, each TS prompt was binary-coded as 0 (disagreement) or 1 (agreement). The reason for that was to avoid magnifying the correlation by reevaluating only those patients within the range of TS. If all the indications were reassessed with a value of 0 (100% disagreement), the patient's score (initially <3) was increased by one point. If no prompt was provided in the first instance, and in the second evaluation, it was thought that the interviewee should have received at least one prompt, the patient's score (initially = 3) was decreased in one point].

#### **2.3. Ethical statements**

This study was performed pursuant to the ethical standards established in the 1964 Declaration of Helsinki. The participants or the patient's caregivers gave their written informed consent and the approval of the Research and Ethics Committee of the Cordoba Hospital was obtained. The neuropsychological evaluation did not pose any risk to the participants who, in all cases, were alert, and willing to perform the complete battery of tests, independently of their relative capacity or willingness to perform some of the subtests in particular. Participants did not receive any payment for their contribution.

The indicators of number of words in the tasks of either spoken or written verbal fluency were

Neurocognitive Implications of Tangential Speech in Patients with Focal Brain Damage

http://dx.doi.org/10.5772/intechopen.71904

201

The relationship between TS and the report of the caregiver during the initial interview, on the one hand, and the complementary behavioral observations during the administration of the battery, on the other hand, were studied through the Spearman's rank-order coefficient (r) for

Additional data were searched with the purpose of discovering the nature of the cognitive impairments associated with TS. Explicitly, if TSP impairment was verified for GNP in general, and for CR or E&T in particular, further analyses on the individual indicators of the GNP component that produced a significant difference between TSP and both non-TSP and HP were performed, thus trying to see the qualitative pattern of TS impairments. MANOVA with TS as grouping variable and the individual indicators of the pertinent GNP components as dependent variables was performed, using Bonferroni post-hoc test for pairwise comparisons. Similarly, if the representative measure of DP produced a significant difference between TSP and both non-TSP and HP, further analyses on its individual indicators were also performed with the same purpose. MANOVA with TS as grouping variable and the individual indicators of DP as dependent variables was performed, using Bonferroni post-hoc test for

In order to outline a hypothetical pattern of cognitive impairments associated with TS all the statistical analyses, including the complementary ones, were taken into account. If some of those cognitive impairments were coincident with measures which have been previously reported as valid indicators of DP, that coincidence was taken as evidence of the validity of the present TS scale. Additionally, inter-rater reliability was analyzed by the intra-class correlation coefficient (ICC). The difference between both evaluations was analyzed by the

A total of 15 cases with a value different from 3 in the TS scale were observed. Only one case was observed with a score of 0 and four cases with a score of 2. Due to such small number of cases, and in order to get better inferences, TS was recoded using 0 when the symptom was

**Table 2** shows that there were no significant differences on type of lesion between non-TSP and TSP. Malignant tumors represented the most frequent type of lesion. By grouping the cells with fewer cases (i.e., the cells with the rest of the lesions), a non-significant difference

absent (non-TSP) and 1 (TSP) when the symptom was present (prevalence 16%).

**Table 1** shows that TSP, non-TSP and HP did not differ in their demographic data.

analyzed by ANOVA, using the Bonferroni post-hoc test for pairwise comparisons.

for dichotomous ones.

ordinal scales or by χ<sup>2</sup>

pairwise comparisons.

Wilcoxon paired-sample test.

**3. Results**

**3.1. Main outcomes**

*2.4.1. Complementary statistical information*

#### **2.4. Statistical analysis**

Demographic data were analyzed by ANOVA for Age (TS as grouping variable) or by Chi square (χ<sup>2</sup> ) for education (three levels: 1st level: primary school, 2nd level: high-school and 3rd level: college or superior) and gender (two levels: men and women).

If the obtained number of TS cases was so small that the presence of empty cells and/or lack of variance could be observed, the original scale of TS was planned to be recoded. Under this condition, and unless otherwise indicated, the groups representing TS were patients without TS (non-TSP), patients with TS (TSP), as well as HP.

The effect of TS on neuropsychological performance was analyzed. With this purpose, a representative measure of the performance in the comprehensive battery was searched by multiplying the errors and times by (−1), and by studying the internal consistency of all the individual indicators through the Cronbach alpha coefficient. If the Cronbach alpha coefficient was satisfactory (0.70 or greater), the individual indicators were added thus obtaining a representative measure of the general neuropsychological performance (GNP). This variable was analyzed by ANOVA with TS as grouping variable and the Bonferroni post-hoc test for pairwise comparisons.3 The possibility of selecting representative measures of both CR and E&T was also analyzed. If that possibility was viable, a bivariate MANOVA with TS as grouping variable and CR and E&T as univariate dependent variables was carried out, using the Bonferroni post-hoc test for pairwise comparisons. The possibility of selecting a representative measure of DP was also analyzed. If that possibility was viable, the individual indicators of the tasks of the battery related with DP were added, and this variable was analyzed by ANOVA with TS as grouping variable and the Bonferroni post-hoc test for pairwise comparisons. If the ANOVA indicates both a significant main effect of TS, and significant pairwise comparisons among the three groups, the association between TS and DP was also analyzed as a way to contribute to the study of the validity of the TS scale (see below). With this purpose, the association by cross tabulation was studied and the percentile partition with the highest χ<sup>2</sup> was reported.

<sup>3</sup> ANOVA is a statistical dependency test. Each significant difference implicates a significant correlation between the independent and dependent variables. In this work no causal relationship between TS and the neuropsychological performance was assumed. The relationship between both variables was studied by ANOVA and, whenever a significant effect was reported, a double implication between the two variables was implicit. Such relationship was always emphasized in the text.

The indicators of number of words in the tasks of either spoken or written verbal fluency were analyzed by ANOVA, using the Bonferroni post-hoc test for pairwise comparisons.

The relationship between TS and the report of the caregiver during the initial interview, on the one hand, and the complementary behavioral observations during the administration of the battery, on the other hand, were studied through the Spearman's rank-order coefficient (r) for ordinal scales or by χ<sup>2</sup> for dichotomous ones.

#### *2.4.1. Complementary statistical information*

**2.3. Ethical statements**

200 Gerontology

**2.4. Statistical analysis**

test for pairwise comparisons.3

was reported.

the highest χ<sup>2</sup>

sized in the text.

3

square (χ<sup>2</sup>

receive any payment for their contribution.

TS (non-TSP), patients with TS (TSP), as well as HP.

This study was performed pursuant to the ethical standards established in the 1964 Declaration of Helsinki. The participants or the patient's caregivers gave their written informed consent and the approval of the Research and Ethics Committee of the Cordoba Hospital was obtained. The neuropsychological evaluation did not pose any risk to the participants who, in all cases, were alert, and willing to perform the complete battery of tests, independently of their relative capacity or willingness to perform some of the subtests in particular. Participants did not

Demographic data were analyzed by ANOVA for Age (TS as grouping variable) or by Chi

If the obtained number of TS cases was so small that the presence of empty cells and/or lack of variance could be observed, the original scale of TS was planned to be recoded. Under this condition, and unless otherwise indicated, the groups representing TS were patients without

The effect of TS on neuropsychological performance was analyzed. With this purpose, a representative measure of the performance in the comprehensive battery was searched by multiplying the errors and times by (−1), and by studying the internal consistency of all the individual indicators through the Cronbach alpha coefficient. If the Cronbach alpha coefficient was satisfactory (0.70 or greater), the individual indicators were added thus obtaining a representative measure of the general neuropsychological performance (GNP). This variable was analyzed by ANOVA with TS as grouping variable and the Bonferroni post-hoc

CR and E&T was also analyzed. If that possibility was viable, a bivariate MANOVA with TS as grouping variable and CR and E&T as univariate dependent variables was carried out, using the Bonferroni post-hoc test for pairwise comparisons. The possibility of selecting a representative measure of DP was also analyzed. If that possibility was viable, the individual indicators of the tasks of the battery related with DP were added, and this variable was analyzed by ANOVA with TS as grouping variable and the Bonferroni post-hoc test for pairwise comparisons. If the ANOVA indicates both a significant main effect of TS, and significant pairwise comparisons among the three groups, the association between TS and DP was also analyzed as a way to contribute to the study of the validity of the TS scale (see below). With this purpose, the association by cross tabulation was studied and the percentile partition with

ANOVA is a statistical dependency test. Each significant difference implicates a significant correlation between the independent and dependent variables. In this work no causal relationship between TS and the neuropsychological performance was assumed. The relationship between both variables was studied by ANOVA and, whenever a significant effect was reported, a double implication between the two variables was implicit. Such relationship was always empha-

3rd level: college or superior) and gender (two levels: men and women).

) for education (three levels: 1st level: primary school, 2nd level: high-school and

The possibility of selecting representative measures of both

Additional data were searched with the purpose of discovering the nature of the cognitive impairments associated with TS. Explicitly, if TSP impairment was verified for GNP in general, and for CR or E&T in particular, further analyses on the individual indicators of the GNP component that produced a significant difference between TSP and both non-TSP and HP were performed, thus trying to see the qualitative pattern of TS impairments. MANOVA with TS as grouping variable and the individual indicators of the pertinent GNP components as dependent variables was performed, using Bonferroni post-hoc test for pairwise comparisons. Similarly, if the representative measure of DP produced a significant difference between TSP and both non-TSP and HP, further analyses on its individual indicators were also performed with the same purpose. MANOVA with TS as grouping variable and the individual indicators of DP as dependent variables was performed, using Bonferroni post-hoc test for pairwise comparisons.

In order to outline a hypothetical pattern of cognitive impairments associated with TS all the statistical analyses, including the complementary ones, were taken into account. If some of those cognitive impairments were coincident with measures which have been previously reported as valid indicators of DP, that coincidence was taken as evidence of the validity of the present TS scale. Additionally, inter-rater reliability was analyzed by the intra-class correlation coefficient (ICC). The difference between both evaluations was analyzed by the Wilcoxon paired-sample test.
