Author : Britta Sophia Redeker
Student ID : 10003872
External supervisor : F.A. Jonker (Clinical Neuropsychologist)
First assessor : Dr. S.P. van der Werf (Clinical Neuropsychologist) Second assessor : Dr. G.H. Tamminga
Research centre / location : Vesalius Altrecht, Woerden
Department, university : Brain and Cognition, University of Amsterdam
Year : 2017
Number of words : 10.517
Index
UNIVERSITY OF AMSTERDAM
Behavioural- versus Conventional
Neuropsychological Measures: Sensitivity and
Specificity after Acquired Frontal Brain Damage
Master thesis Healthcare Psychology - Clinical Neuropsychology
2
1. Abstract 3
2. Introduction 4-8
2.1 Behavioural assessment of frontal lobe dysfunction 5
2.2 Social cognition 6 2.3 Dissociation between cognition and behaviour 6
2.4 Relevance and purpose of this study 8
3. Methods 9-17 3.1 Sample characteristics 9 3.2 Procedure 9 3.3 Materials 11 3.4 Data analysis 16 3.4.1 Preliminary analysis 16
3.4.2. Comparing cognition and behaviour: FBD versus NBD 17
3.4.3 The relationship between cognition and behaviour 17
4. Results 18-21
4.1 Differences in cognition and behaviour: FBD versus NBD 18
4.1.1 Cognitive executive functioning 18
4.1.2 Social cognition 19
4.1.3 Behaviour 19
4.2 Relationship between cognition and behaviour 20
4.2.1 Behaviour and cognitive executive functioning 20
4.2.2 Behaviour and social cognition 21
5. Discussion 21-25
5.1 Conventional neuropsychological measures of cognitive executive functions 21
5.2 Social cognition 22
5.3 Frontal Systems Behaviour Scale (FrSBe) 23
5.4 Dissociation between cognition and behaviour 23
5.5 Limitations 24
5.6 Conclusion 25
3
1. Abstract
This cross-sectional study examined the sensitivity and specificity of traditional neuropsychological measures of cognitive executive functioning, social cognition and the behavioural questionnaire Frontal Systems Behaviour Scale (FrSBe family rating version; Grace & Malloy, 2001), first by comparing two groups of patients with cognitive complaints: a group with objectified frontal brain damage (FBD) in the chronic phase after acquired brain injury (ABI; N = 35; years since lesion M = 13.6) and a group without structural brain damage (NBD; N = 28;
years since accident M = 11.5), and second by investigating the relationship between cognitive
performance and three behavioural syndromes: Apathy, Disinhibition and Executive functioning (i.e. FrSBe subscales). Brain damage was confirmed by 3 Tesla MRI scans and the expert opinion of independent radiologists. Using independent samples t-tests and Mann-Whitney tests to
compare cognitive performance and behaviour between the two groups, no significant and / or clinically relevant differences in performance on most of the traditional measures of cognitive executive functioning, social cognition, and behaviour were found. Although the FBD group scored significantly lower on a test of verbal fluency than the NBD group, this result was not clinically relevant, as can be seen as a trend in the data. Using regression analysis no significant associations between behaviour and cognition were found, indicating a dissociation between cognition and behaviour. Moreover, both groups scored – on average and according to norm tables – within the normal range on conventional tests, but within the clinically significant high range on FrSBe’s behavioural scales. Together these results strongly suggest a lack of sensitivity and specificity of conventional neuropsychological measures in a neuropsychiatric population with acquired FBD and stress the importance of developing and validating new methods with higher ecological validity.
Key words: sensitivity, specificity, dissociation, neuropsychiatry, acquired brain injury (ABI), frontal
brain damage (FBD), no brain damage (NBD), social cognition, cognitive executive functioning, behaviour, apathy, disinhibition, executive dysfunction, Frontal Systems Behaviour Scale (FrSBe)
4 2. Introduction
‘The tamping iron entered the cranium, passing through the anterior left lobe of the cerebrum, and made its exit in the medial line, at the junction of the coronal and sagittal sutures, lacerating the longitudinal sinus, fracturing the parietal and frontal bones extensively, breaking up considerable portions of the brain, and protruding the globe of the left eye from its socket, by nearly half its diameter.’
-Harlow, 1848
With his famous report on Phineas Gage, Harlow made an important contribution to the theories of ‘phrenology’ or so called ‘cerebral localization’ in the nineteenth century. Although Gage physically recovered - retaining normal intelligence, language and memory functions - his mind and behaviour were radically changed: ‘The equilibrium or balance, so to speak, between his intellectual faculties and animal propensities, seems to have been destroyed.’ (Harlow, 1868). Nowadays, the
relationship between personality and behavioural change is still the most evident following damage to the frontal lobes.
In the years since Harlow’s report, a growing number of studies have tried to clarify the role of the frontal lobes in human cognition and behaviour. Damage to the frontal lobes has been related to personality change and severe social-behavioural disturbances, including disinhibition, impulsivity, distractibility, perseveration, stereotyped behaviour, utilization behaviour, emotional lability, apathy, lack of tact, disregard of social conventions, as well as cognitive problems, such as: executive dysfunctions, working memory problems, attentional problems, mental inflexibility, and decreased abstract thinking ability (Barrash, Asp, Markon, Manzel, Anderson, & Tranel, 2011; Stuss & Alexander, 2000; Harlow, 1848; Eslinger & Damasio, 1985). These changes can seriously restrict a patients’ personal autonomy and socio-economic status, decreasing overall quality of life and increasing family burden. Subsequent psychological distress can lead to mood disorders, such as depression and increased suicide risk (Fleminger, Oliver, Williams, & Evans, 2003).
Although in traditional neuropsychology it is often assumed that specific cognitive disorders, especially executive dysfunctions, are responsible for the behavioural changes following frontal brain damage (Ready, Stierman, & Paulsen, 2010; Ardila, 2008), research has suggested a dissociation between cognition and behaviour in patients with frontal lesions (Slachevsky, Peña, Pérez, Bravo, & Alegría, 2006; Namiki et al., 2008; Knutson et al., 2015;
5
Funayama, Mimura, Koshibe, & Kato, 2010). Whereas cognition has been the main focus in most neuropsychological oriented studies - resulting in various instruments for the assessment of cognitive disturbances - fewer methods are available for the assessment of behavioural disturbances following frontal brain damage (Lezak, 2004), notwithstanding growing evidence that these behavioural methods may be more sensitive to the changes following frontal brain damage than traditional neuropsychological tests (Namiki et al., 2008; Knutson et al., 2015).
2.1 Behavioural assessment of frontal lobe dysfunction
Compared to conventional neuropsychological measures of frontal lobe functioning, behavioural measures may be more sensitive to the changes following damage to the frontal lobes (Slachevsky et al., 2006; Namiki et al., 2008; Knutson et al., 2015; Caracuel, et al. 2008). Examples of behavioural instruments frequently used in clinical practice are the Neuropsychiatric Inventory (NPI; Cummings et al., 2004), the Apathy Evaluation Scale (AES; Marin, Biedrzycki, & Firincioguliari, 1991), and the Frontal Systems Behaviour Scale (FrSBe; Grace & Malloy, 2001; formerly known as the Frontal Lobe Personality Scale (FLOPS), Grace, Stout, & Malloy, 1999). The FrSBe is of particular interest in analysing behavioural disturbances related to frontal brain damage, and is a validated instrument for the assessment of three behavioural syndromes: Executive dysfunction, Apathy, and Disinhibition. These syndromes are theoretically related to damage to specific frontal-subcortical neural circuits, respectively: 1) a dorsolateral prefrontal circuit, 2) an anterior cingulate (ACC) and superior frontal circuit, and 3) an inferior medial frontal circuit (Cummings, 1993; Mega & Cummings, 2007; Alexander, Delong, & Strick, 1986; Carvalho, Ready, Malloy, & Grace, 2013; Stout, et al., 2003; Bonelli & Cummings, 2007). According to literature, the dorsolateral prefrontal cortex (DLPFC) is most involved with the executive functions (e.g. planning, attentional control, working memory, cognitive flexibility, reasoning, problem solving and monitoring of behaviour; Ardila, 2008; Damasio & Eslinger, 1985; Stuss & Benson, 1984). Damage to this area should in theory result in a dysexecutive syndrome, characterized by disorganized behaviour, perseveration, impaired reasoning, and mental inflexibility (Mega & Cummings, 2007). The anterior cingulate and superior frontal circuit is most involved in initiation and motivated behaviour. Damage to this region has been associated with apathy, abulia, and psychomotor retardation (Cummings & Bonelli, 2007). The inferior medial frontal circuit, including OFC and ventromedial prefrontal cortex (vMPFC), is closely connected to subcortical structures (i.e. the limbic system, hypothalamic, amygdale, hippocampal and brainstem regions; van Hoesen, Pandya, & Butters, 1975; Krueger, Barbey, & Grafman, 2009), and is assumed to mediate socially appropriate behaviour by controlling
6
impulses from these subcortical structures (Blair, 2004). Damage to the inferior frontal regions has been associated with severe social-behavioural problems (Eslinger & Damasio, 1985; Leopold et al., 2012; Gilbert et al., 2006) and emotional disturbances (Shamay-Tsoory et al., 2003), characterized by social inappropriateness such as a lack of tact and disregard of social conventions (Bachevalier & Loveland, 2006; Blair, & Cipolotti, 2000; Hornak, Rolls, & Wade,
1996; Rolls, Hornak, Wade, & McGrath, 1994), disinhibition (Berlin et al., 2004; Knutson et al.,
2015; Namiki et al., 2008), poor risk assessment and decision making (Fellows & Farah, 2003; Funayama, Mimura, Koshibe, & Kato, 2010).
In terms of clinical relevance, the FrSBe has been shown to reliably distinguish patients with frontal lobe damage from normal controls, and addicted patients (Caracuel, et al. 2008). Regarding ecological validity, it has been a useful instrument in predicting community integration (Reid-Arndt, Nehl, & Hinkebein, 2007; ) and engagement in risky and aggressive behaviours (Ready et al., 2001).
2.2 Social cognition
It has been suggested that social cognitive tests are more sensitive to frontal brain damage (Mitchell, Avny, & Blair, 2006; Namiki et al., 2008; Hornak et al., 1996; Blair & Cipollloti, 2000). For example, Mitchell et al. (2006) found that patients with lesions involving ventromedial PFC were significantly impaired in affective Theory of Mind (ToM), i.e. they performed worse in terms of identification and estimating the severity of socially inappropriate behaviour than a control group without brain damage. Some researchers have even suggested a relationship between behavioural disinhibition and performance on social cognitive tests (e.g. the Strategic Emotional Intelligence items on the Mayer-Salovey-Caruso Emotional Intelligence Test; Balanced Emotional Empathy Scale; Facial Expression Recognition Task) in patients with inferior frontal brain damage (Namiki et al., 2008; Leopold et al., 2012; Hornak et al., 1996; Blair & Cipollloti, 2000). Namiki et al. (2008) concluded that the disinhibitory behaviour seen in patients with OFC damage may result from the inability to interpret social signals and facial emotions. However, the empirical value of these case studies is limited and other studies have proposed alternative interpretation. Berlin, Rolls and Kischka (2004) stated that the inappropriate social behaviour of patients with frontal brain damage results from insensitivity to reward and punishment rather than the inability to interpret emotions. Consequently, the relationship between social cognition and behaviour remains unclear and needs to be further investigated.
7
Although there seems to be a relationship between social cognition and behaviour, patients with damage involving inferior frontal and orbitofrontal cortex (OFC) have been reported to display severe social-behavioural problems - often characterized by disinhibition and risk-taking behaviour - despite remaining intellectually and cognitively intact on conventional intelligence and neuropsychological tests frequently used in clinical practice (Eslinger & Damasio, 1985; Ready, Stierman, & Paulsen, 2010), such as the Wechsler Adult Intelligence Scale III / Revised (WAIS-R; Knutson et al., 2015; Namiki et al., 2008; WAIS-III, Funayama et al., 2010), Wechsler Memory Scale Revised (WMS-R), Behavioural Assessment of the Dysexeutive Syndrome (BADS), the Wisconsin Card Sorting Test (WCST) (Namiki et al., 2008; Funayama et al., 2010), Trail Making Test (TMT), and the Mini Mental State Examination (MMSE; Funayama et al., 2010). Using only these test results, FBD patients may be wrongly diagnosed with malingering, or behavioural symptoms may be explained in psychiatric terms, complicating subsequent treatment choice and thereby treatment efficacy. Since the frontal cortex is assumed to play a major role in generating flexible and adaptive (social-)behaviour based on a complex interplay between emotions, expectations and changing environmental cues, it has been stated that the conventional and rather deterministic neuropsychological measures do not encompass the dynamic properties of daily life situations, and may therefore lack sensitivity to the changes following frontal brain damage (Slachevsky et al., 2006; Ardila, 2008). Furthermore, since most of the current traditional measures of ‘frontal lobe function’ (e.g. TMT, Stroop Test, Fluency tests, WCST) are multifactorial in nature, performance on these tests relies on a widespread and closely interconnected neural network, implying that dysfunction in one region may affect functions in other regions (Ardila, 2008; Alvarez & Emory, 2006). Although these tests may be more sensitive to damage involving superior frontal regions, performance on these tests seems not specifically related to frontal lobe functioning (Alvarez & Emory, 2006).
Whereas previous studies concerning damage to inferior frontal regions found a dissociation between cognition and behaviour (Namiki et al., 2008;), studies concerning damage to superior frontal regions have shown conflicting results regarding this relationship (Anderson & Bergedalen, 2002; Schiehser et al., 2011). For example, self-reported depressive symptoms rather than self-reported behavioural executive dysfunction have been related to performance on neuropsychological tests of executive function (Shiehser et al., 2011). Although DLPFC damage has previously been associated with worse performance on neuropsychological tests of executive functioning (Alvarez & Emory, 2006), this now seems at least partly independent from self-reported behavioural executive dysfunction (Anderson & Bergedalen, 2002; Schiehser et al., 2011). Regarding apathy, cognitive performance seems to be only related to the cognitive aspects
8
of apathy and not to the behavioural or emotional aspects of the construct (Andersson & Bergedalen, 2002). Since the FrSBe particularly measures behaviour, it is possible that its measured constructs are at least partly independent of cognitive dysfunction (Stout, Wyman, Peavy, & Salmon, 2003; Schiehser et al., 2011; Namiki et al., 2008).
2.4 Relevance and purpose of this study
Contrary to growing evidence for FrSBe’s proposed brain-behaviour relationships, just a few studies – to our knowledge – have directly related behavioural change to neuropsychological outcome in a chronic outpatient group with objectified frontal ABI. Studies suggesting moderate correlations between cognition and behaviour seem to be largely restricted to populations with neurodegenerative diseases such as frontal dementia and multiple sclerosis, lack well-defined lesions, or report single cases. Moreover, most studies used self-reported information to classify severity of TBI (e.g. loss of consciousness, initial Glasgow Coma Scale (GCS), and duration of posttraumatic amnesia (PTA)), instead of using imaging material to objectify lesion location. This may have confounded the purity of the target population. Consequently, the empirical value of studies involving patients with frontal ABI is restricted.
With regard to clinical relevance, it is of particular importance to explore the relationship between cognition and behaviour in patients with frontal lobe damage, since the earlier described conventional neuropsychological tests of ‘frontal lobe functioning’ are still used in clinical practice. They are interpreted in terms of causal relationships between cognitive dysfunction and behavioural problems. Clinicians might be drawing conclusions based on false assumptions, thereby ignoring other causes for behavioural distortion, which may have major implications for treatment choice and thereby treatment efficacy.
The purpose of this study is to investigate whether the FrSBe is more sensitive and specific to the changes following frontal brain damage than conventional tests. First, by comparing neuropsychiatric outcome between patients with objectified FBD and patients without structural brain damage, and second, by studying the relationship between cognition and behaviour in both groups. It is hypothesized: that the FrSBe has a higher sensitivity for (clinically) significant changes following FBD than conventional tests of cognitive executive functions. FrSBe’s subscales (Apathy, Disinhibition, and Executive Dysfunction) are - at least partly - independent of performance on conventional tests of executive functions: 1) behavioural executive dysfunction is at most moderately related to performance on conventional tests of executive functions (Schieser et al., 2011; Paulsen et al., 2000; Alvarez & Emory, 2006), 2) behavioural apathy and behavioural disinhibition are not related to performance on conventional
9
tests of executive function (Schieser et al., 2011; Paulsen et al., 2000; Namiki et al., 2008; Funuyama et al., 2010), and 3) behavioural disinhibition is related to social cognitive problems in patients with FBD (Namiki et al., 2008; Leopold et al., 2012; Hornak et al., 1996; Blair & Cipollloti, 2000).
3. Methods 3.1 Procedure
Permission for this study was provided by the Ethics Committee of Altrecht. Demographic characteristics (age, sex, educational level, time since accident) were collected by a neuropsychologist during intake sessions. All patients completed a battery of clinically validated neuropsychological tests, social cognitive tests, and behavioural questionnaires. SO’s completed the questionnaires separately. Both patients and SO’s signed an informed consent before using their data in this study.
3.2 Sample Characteristics
A total of 63 patients with cognitive complaints, who were part of the diagnostic trajectory at
Altrecht Vesalius, were selected. Vesalius is an outpatient mental health centre specialized in neuropsychiatry and ABI. Inclusion criteria were: cognitive complaints, structural frontal brain damage (confirmed with 3-tesla MRI scans) due to trauma, stroke or tumours, or no structural brain damage (confirmed with 3-tesla MRI scans) and no significant loss of consciousness, and had a significant other (SO; i.e. someone who knows the patient well, preferably before and after the ABI) to complete the FrSBe. Due to the scarcity of patients with frontal brain damage, this study also included patients with lesions in additional lobes, such as parietal and temporal and used data from the Vesalius database. This database contains data from patients who already had a neuropsychological assessment at Vesalius in the past (2008 - 2016) and had given their written informed consent. Patients were excluded if they were not able to reliably complete the test battery, had disturbed vision, insufficient knowledge of the Dutch language, had a history of substance abuse / intoxication, mental retardation, pervasive developmental disorder (according to the Diagnostic and Statistical Manual of Mental Disorders–Fourth Edition, DSM–IV), or neurological
disorder, including those due to neurodegenerative disease, infectious disease, metabolic disease or other diseases known to affect the central nervous system. Patients were divided into two groups: 1) a frontal brain damaged group (FBD; N = 35) and 2) a control group without
10
consent. Educational level was rated according to the Verhage Scale (Verhage, 1964). This scale consists of seven categories: 1) less than six years of primary education, 2) six years of primary education (finished), 3) six years of primary education and less than two years of low level secondary education, 4) four years of low level secondary school, 5) four years of average level secondary education, 6) five years of high level secondary education, and 7) university degree. Sample characteristics are presented in table 1, page 10.
Preliminary analysis showed no significant differences in mean age and time since accident / lesion between the two groups. Chi-squared tests showed no significant differences in educational level between the two groups. There was however, a significant difference in sex distribution, with more males in the FBD group and more females in the NBD group, which is a realistic representation of the ABI population. Mean educational levels for both patient groups and SO’s corresponded to Verhage code 5, indicating four years of average level secondary education.
Table 1. Demographics of the sample and injury related variables
Total
(N = 63) FBD (N = 35) NBD (N = 28) FBD vs
NBD
Mean (SD) Mean (SD) Mean (SD) T-score p-value
Patients
Age (years)
Time since lesion / accident (years)
Educational degree1
Significant others (SO)
Age (years) Educational degree1 41.60 (12.15) 12.69 (13.63) 5.24 (1.16) 49.82 (11.59) 5.33 (1.05) 42.25 (12.56) 13.55 (13.67) 5.11(1.16) 51.53(12.41) 5.10 (1.16) 40.80 (11.80) 11.53 (13.77) 5.39 (1.17) 47.79 (10.40) 5.59 (0.84) 0.47 0.56 χ² 7.02 T-score 1.24 χ² 4.77 0.642 0.580 0.219 0.221 0.445 Frequency N (%) Frequency (%) N χ² Patients Male Female
Significant others (SO)
Male Female 38 (60.3%) 25 (39.7%) 20 (31.7%) 40 (63.5%) 26 (74.3%) 9 (25/7%) 9 (25.7%) 24 (68.6%) 12 (42.9%) 16 (57.1%) 11 (39.3%) 16 (57.1%) 6.42 1.21 0.019* 0.288
11 Aetiology of lesion2 Traumatic Vascular3 Tumours4 Frontal+ Pure frontal 23 (65.7%) 11 (31.4%) 4 (11.4%) 21 (60%) 14 (40%)
* Significance level set at p < .05, two-tailed
1 According to the Verhage Scale
2 The brain injury can be a consequence of multiple aetiologies (traumatic and vascular: N = 3)
3 Also acquired epilepsy (N = 2)
4 Also acquired epilepsy (N = 2)
Depression
There was no significant difference in average BDI-II scores the groups, t(57) = -1.043, p >.05.
Both groups scored within the moderate range, indicating moderate depression (table 2 page 11) .
Table 2. Depressive symptoms - Beck Depression Inventory II (BDI-II)
FBD NBD
Mean (SD) Mean (SD) t df p r
BDI-II 20.09 (11.13) 23.26 (12.16) -1.043 57 0.301 0.14
Independent samples t-test
Significance level set at p < .05, two-tailed; clinical relevance level set at p < 0.006 (Bonferroni correction)
Small effect size: 0.1; medium effect size: 0.3; large effect size: 0.5
3.3 Materials Lesion analysis
Since high-resolution Magnetic Resonance Imaging (MRI) techniques (at least 3 tesla) have shown to be more sensitive to small frontal lesions than standard MRI techniques or CT-scans (Namiki et al., 2008), this study used imaging data from 3 tesla MRI scans. Brain damage was objectified by independent radiologists in a clinical setting, who were blind to the purpose of this study.
Neuropsychological assessment Executive functioning
Executive functioning was measured with the Wisconsin Card Sorting Test (WCST), Letter Fluency (DAT), Trail Making Test (TMT), and the Stroop Colour-Word Interference Task (Stroop). The psychometric properties and a description of these tests are described below.
12
The WCST assesses cognitive flexibility (i.e. the ability to shift cognitive strategies in response to changing environmental contingencies) and abstraction ability (Straus, Sherman, & Spreen, 2006). Participants are asked to sort cards according to their attributes shared with four stimulus cards which differ in three perceptual dimensions (colour, shape, and number of objects). The sorting rule is not specified - participants are only told if a particular card sorting is correct or incorrect - and changes without warning after ten correct sorting’s. The test takes approximately 10-20 minutes and generates different scores, i.e. number of categories completed, number of perseverative and non-perseverative errors and answers, and number of answers on a conceptual level. This study used ‘perseverative errors’ (WCST-PE) for data analysis (Robinson, Heaton, Lehman, & Stilson, 1980).
The reliability estimates range from very low (e.g. stability coefficients ≤ .12 ) to moderate (e.g. generalizability coefficients .37 to .72) and excellent (e.g. interrater reliability > .83) (Strauss, Sherman, & Spreen, 2006). Overall the reliability seems to be higher in clinical samples (e.g. perseverative errors) than in nonclinical samples (Tate, Perdices, & Maggiotto, 1998). The WCST is sensitive but not specific to dorsolateral prefrontal brain damage (MacPherson et al., 2002). This means that performance on this test cannot predict focal FBD, since any significant impairment in other processes (e.g. deficits in visual processing, attention, working memory, etc.) may lead to poor performance on the WCST (Strauss, Sherman, & Spreen, 2006). Furthermore, depression, anxiety and mental fatigue may influence WCST performance, resulting in increased perseveration, failure to maintain set, and decreased conceptual level (Moritz et al., 2002; Toren et al., 2000; van der Linden et al., 2003; Strauss, Sherman, & Spreen, 2006). Although the WCST requires numerous skills, shifting ability seems to contribute significantly to WCST performance (Miyake, et al., 2000). Poor performance on the WCST has been associated with caregivers’ reports on behaviour and cognitive deficits in a patient group (Burgess et al., 1998).
DAT Letter Fluency (norms: 17 - 89 years)
The verbal phonetic fluency test (DAT) measures verbal functioning and executive control. In this task executive control can be described in terms of finding and using strategies to access lexicon, attention / focus, selection based on certain task-specific constraints, and inhibition of inappropriate responses (Schmand, Groenink, & van den Dungen, 2008). In three trials participants are asked to produce as many words in a given time (60 seconds), starting with the letters D, A, and T, while following three rules: 1) no words starting with a capital letter (e.g. names and places), 2) no numbers, and 3) no subsequent words starting with the same prefix. Normative data is derived from a healthy population (N = 200), and standard scores can be
13
calculated based on age, educational level and verbal intelligence (Nederlandse Leestest voor Volwassenen, NLV; Schmand, Bakker, Saan, & Louman, 1991). Data analysis was carried out for the total score (total number of words).
The internal reliability and parallel-test reliability of the DAT are good (respectively .82 and .78; Schmand et al., 2008). Concerning the construct validity of this test, correlations with category fluency are relatively high (r =.37 to r =.50) compared to Ruff Figural Fluency Test (r
=.19), Stroop task (r =.33), a verbal vocabulary test (NLV; r =.37) and the Boston Naming Test (r =.19). According to Schmand et al. (2008) performance on this test is only influenced by
educational level and / or vocabulary.
Stroop (norms: 8-65 years)
This test measures selective attention, sensitivity to interference and inhibition of automated responses. It consists of three stimulus cards: 1) a colour-naming card with squares printed in different colours, 2) a word-reading card with colour words printed in black ink, and 3) a colour-word-naming card with colour words printed in incongruent colours. The third card requires participants to name the colour of the letters (independently of the written words) as quickly as possible, and generates an interference score. The present study used the ratio score for data analysis, which was calculated by dividing completion time of the Inhibition card by the time required to complete the Reading card.
Reliable assessment requires an intact visuomotor system. The psychometric properties of the test meet all COTAN criteria (ranging from sufficient to good), except for the criterion validity, which is insufficient. The test-retest reliability is high (colour-naming card: r = .87; word-naming card: r = .91; colour-word-word-naming card: r = .89; Bouma, Mulder, Lindeboom, Mulder, & Schmand, 2012). Different comprehensively researched Dutch norms are available (van der Elst, van Boxtel, van Breukelen, & Jolles, 2006; Schmand, Houx, & de Koning, 2003; Strauss, Sherman, & Spreen, 2006; Lezak, Howieson, & Loring, 2004; Bouma & Lindeboom, 1996).
TMT (norms: >11 years)
The purpose of this test is to measure cognitive flexibility and divided attention. Since it is a multifactorial test relying on different cognitive abilities, it also measures visual searching and scanning, sequencing, psychomotor speed, abstraction and working memory (Straus et al., 2006). It consists of two parts: A (number sequencing) and B (number-letter switching). Part A requires participants to draw lines between circles (connecting numbers in ascending order) as quickly as possible. Part B requires participants to draw lines to connect circles while alternating between
14
numbers and letters (i.e. 1-A-2-B-3-C, etc.). The test provides standardized scores for both completion times and error types (e.g. sequential errors and set-loss errors). This study used the ratio score for the data analysis, which was calculated by dividing the completion time for the number-letter sequencing card, by the time on the number sequencing card.
According to COTAN criteria (1992), the psychometric properties of this test are insufficient. More recent research has shown moderate reliability for part A (r = .70 - .79) and
moderate to high reliability for part B (r = .70-.89) (Bouma et al., 2012). Norms are also available
for patients with ABI (Kok & van Dam, 2012).
Social cognition
Social cognition will be assessed with the Faux Pas test (affective ToM) and Reading the Mind in the Eyes Test (emotion recognition).
Faux Pas
This test consists of four stories with a faux pas (i.e. a story in which the speaker unintentionally says something hurtful or insulting to the listener), and four stories without a faux pas. The stories are read aloud by the researcher, while the patients are able to read the stories along (correcting for a memory component). After each story the following questions are asked:
1. Did anyone say something awkward or something they should not have said? (faux pas detection). If the answer is yes, the following questions are asked:
2. Who said something awkward or something he / she shouldn’t have said? 3. Why shouldn’t he/she have said that? (understanding inappropriateness) 4. Why do you think he / she said that? (understanding intentions / motivations)
5. Did he / she (i.e. the person who said something awkward) know / realize that X? (understanding belief)
6. How do you think he / she felt? (empathy).
In order to control for (verbal) story comprehension, two or three control questions are asked. Participants will receive 0 points if they say ‘no’ to the detection question of a faux pas story. They can still receive points for the control questions. The maximum score is 49. This study used the scores of the four faux pas stories (maximum score = 24) and the comprehension questions (maximum score = 8).
The psychometric properties of the test are discussed in Gregory, Lough, Stone, Erzinclioglu, Martin, et al., 2002); they showed that the inter-rater reliability was excellent (r =
15
.98) and that the test can discriminate between Alzheimer’s disease and the frontal variant of frontotemporal dementia.
RMET (norms: 17-35 years)
This test measures the ability to evaluate someone’s emotional state based on pictures of the eye-region (i.e. theory of mind). Participants are presented with a series of 36 pictures, and asked to choose which of four words describes best what the person in the picture is feeling. The maximum score is 36.
Norms are available for adults with Asperger syndrome and a normal population (M =
26.2, SD = 3.6). It has been shown to have good discriminant validity: it distinguishes very high functioning adults with Asperger Syndrome / high functioning autism from controls (Baron-Cohen, Wheelwright, Hill, Raste & Plumb, 2001). It also has been found to be useful in other clinical groups, such as in patients with prefrontal lesions or damage due to amygdalectomy (Stone, Baron-Cohen, & Knight, 1999; Stone, Baron-Cohen, Young, & Calder, 1998). In a Spanish population of undergraduates it has been shown to have good test-retest reliability (.63, p = <.01) (Fernández-Abascal, Cabello, Fernández-Berrocal, & Baron-Cohen, 2013). There are also Dutch norms available (Kan, 2007).
Behavioural measures
FrSBe
Behaviour was measured using the family-rating version of the FrSBe (Grace & Malloy, 2001). This 46-items behaviour rating 5-point likert scale (1 almost never - 5 almost always) provides a total frontal disturbance score and three subscale scores: Apathy (14 items), Disinhibition (15 items) and Executive dysfunction (17 items). These subscales are based on a neurological theoretical framework of frontal lobe function. Higher scores represent greater neurocognitive / behavioural impairment (clinical significance: t-scores ≥ 65; borderline significance: t-scores
60-64). According to the authors (Grace & Malloy, 2001) and factor analysis studies (Stout, Ready, Grace, Malloy, & Paulsen, 2003; Carvalho et al., 2013) the three behavioural clusters (Apathy, Disinhibition, and Executive dysfunction) are the most frequently reported behavioural syndromes following damage to the frontal lobes and can be associated with specific regional frontal lobe disturbances (DLPFC: executive dysfunction; ACC: apathy; OFC: disinhibition). These studies provide evidence for high internal consistency, strong validity and reliability of the scales (Total score: α =.89; Apathy: α = .89; Executive dysfunction: α = .85; Disinhibition: α = .90). Stout et al. (2002) showed that the three behavioural clusters of the FrSBe account for
16
40.7% of the common variance among the items. Sixty-three percent of these items loaded only on their hypothesized factor. The intercorrelations between the three subscales range from small (r = .22) to moderate (r = .43; Stout et al., 2003). Test-retest reliability (in the family-rating
version) has been shown to be acceptable (.78; Velligan, Ritch, Sui, DiCocco, & Huntzinger, 2002). In one study, the patient-rating version of the FrSBe did not surpass the traditional cut-off for establishing test-retest reliability (<.70; Niemeier, Perrin, Holcomb, Nersessove, & Rolston, 2013).). The authors stated that only the family-rating versions are useful, since self-awareness in patients with frontal brain damage is often impaired. This study therefore only used the family-rated version of the FrSBe. FrSBe total and subscale scores in the data analysis.
Norms are derived from a community sample of 436 men and women (American), including several clinical groups (patients with frontotemporal dementia, frontal lesions, non-frontal stroke, head injury, Alzheimer’s disease, Huntington’s disease, and Parkinson’s disease), and based on two educational levels (≤ 12 years of education; ≥ 12 years of education) and different age groups (18-39 years; 40-59 years; ≥ 60 years).
Depression
The Beck Depression Inventory II (BDI-II; Beck, Ward, Mendelson, Moch, & Erbaugh, 1961) is a 21-item multiple-choice self-report questionnaire measuring severity of cognitive, behavioural and physical symptoms of depression. Patients are asked to choose one of four statements that most accurately describes how one has felt over the past two weeks. It takes approximately 5-10 minutes to complete the questionnaire. Standardized cut-off points are provided by the manual (Beck et al., 1961). Higher scores indicate more depressive symptoms (score range: 0 – 63): minimal (0-13), mild (14-19), moderate (20-28), and severe (29-63).
3.4 Data analysis
Statistical analysis was conducted using IBM SPSS 23.0 for Windows (IBM Corp., Armonk, New York). For all the analysis described below, a two-sided αlevel of 0.05 was used. Additionally, an α level of 0.1 was used to investigate trends in the data. Non-normally distributed variables (Kolmogorov-Smirnov p <.05; -2 < skewness and kurtosis > 2) were (if possible) transformed
into standard t-scores.
3.4.1 Preliminary analysis
FBD: Preliminary analysis showed that the assumptions of normality (Kolmogorov-Smirnov) were violated for the following variables: Stroop, D(35) = 0.22, p < .001, WCST, D(35) = 0.16 , p
17
< .05, and Faux pas D(35) = 0.17, p <.05. Measures of skewness and kurtosis showed
problematic skewness (2.16) and kurtosis (5.31) for Stroop. The other variables were within the acceptable range between -2 and 2. Transformation of Stroop ratio scores into standard t-scores
per patient solved problems with skewness and kurtosis (> -2 and < 2), but did not solve the normality problem, D(35) = 0.18, p <.05. Therefore, these variables were analysed using
non-parametric methods. WCST standard t-scores showed no violations of normality, D(35) = 0.14, p
= 0.077, or problematic skewness and kurtosis. Thus, these standard scores were used for data analysis.
NBD: Preliminary analysis showed that the assumptions of normality (Kolmogorov-Smirnov) were violated for the following variables: Stroop, D(27) = 0.27, p <.001, Faux pas, D(28) = 0.18, p <.05, and FrSBe Apathy, D(28) = 0.27, p <.001. Measures of skewness and kurtosis showed
problematic skewness (2.55) and kurtosis (7.66) for Stroop and problematic skewness 2.72) and kurtosis (8.82) for FrSBe Apathy. The other variables were within the acceptable range between -2 and -2. After transforming all FrSBe subscale scores into standard t-scores per patient,
preliminary analysis showed no violations of normality, skewness, and kurtosis were within the acceptable range between -2 and 2 (both groups). Therefore, FrSBe standard scores were used for data analysis.
Homogeneity of variance
Levene’s test indicated unequal variances only for the TMT, F = 6.82, p = 0.011; degrees of
freedom were adjusted from 61 to 59.
Outliers
Stem-and-leaf plots showed different extreme values for every variable. Log transformations did not solve this problem; after deleting outliers, new outliers appeared. Therefore, outliers were not deleted for data analysis, which may have influenced the results below.
3.4.2 Comparing cognitive executive functioning, social cognition, and behaviour: FBD versus NBD
For the normally distributed data, independent samples t-tests were conducted to compare the
two patient groups on measures of cognitive executive functioning (TMT, WCST, DAT), social
cognition (RMET), behaviour (FrSBe) and depression (BDI-II). In order to correct for multiple testing, a Bonferroni multiple-significance-test correction was applied by dividing p < .05 by the
18
were set at α level p < .006 in order to reject the null-hypothesis of no effect. Effect sizes were
calculated using Pearson’s correlation coefficient r; clinical relevance was set at r ≥ .3 (medium effect size).
For the non-normally distributed data, Mann-Whitney tests were conducted to compare the two patient groups on measures of cognitive executive functioning (Stroop) and social
cognition (Faux pas). Effect sizes (r) were calculated using z-scores (Rosenthal, 1991).
3.4.3 The relationship between cognition and behaviour
The relationship between behaviour and cognitive executive functioning was analysed using four different multiple regression analysis. In this design the independent variables were cognitive executive functioning (TMT, Stroop ratio, DAT, and WCST perseverative errors) and the dependent variable was behaviour (FrSBe total and subscale scores). Another four different multiple regression analysis were carried out in order to examine the relationship between behaviour and social cognition, by using social cognition (Faux pas, RMET) as independent variables and FrSBe total and subscale scores as dependent variables.
4. Results
4.1 Differences in cognitive executive functioning, social cognition, and behaviour: FBD versus NBD
4.1.1 Cognitive executive functioning
Independent samples t-tests and additional calculations of Pearson correlation coefficients
showed no significant and clinically relevant differences between the groups in mean performance on the TMT and WCST (p > 0.05, r <.30, table 3, page 18). On average, the mean
DAT score was significantly higher in the NBD group than the FBD group, t(61) = -2.06, p =
.043. This result indicates that patients with FBD perform significantly worse on a measure of verbal functioning and executive control (i.e. finding and using strategies to access lexicon). Nevertheless, taking into account the Bonferroni correction for multiple testing (p > .006) and
the found effect size, r = .26 (small-sized), this result was not clinically relevant. A
non-parametric test (Mann-Whitney Test) showed no significant difference between the two groups on the Stroop task (p > .05, table 3, page 17). In summary, these results indicate that FBD
patients in the chronic phase after ABI do not perform significantly different from NBD patients on most traditional neuropsychological measures of executive cognitive functioning often used in clinical practice. Moreover, according to norm tables, both groups scored within the average
19
range on all the tests, indicating no cognitive executive problems. These tests may therefore lack sensitivity and specificity to the changes following structural frontal brain damage.
Table 3. Cognitive executive functioning
FBD NBD
Independent samples t-tests Mean (SD) Mean (SD) t df p r
TMT ratio score DAT total score
WCST PE standard score 2.37 (0.96) 29.43 (10.89) 45.09 (11.97) 2.06 (0.63) 35.86 (13.85) 43.50 (10.49) 1.52 -2.06 0.55 59 61 61 0.135 0.043* 0.583 0.19 0.26 0.07
Mann-Whitney test Mean rank Mean rank U Z p r
Stroop ratio score 31.77 31.15 463.0 -0.14 0.893 0.02
Independent samples t-tests and Mann-Whitney u test
*Significance level set at p < .05, clinical relevance level set at p < 0.006 (Bonferroni correction)
Small effect size: 0.1; medium effect size: 0.3; large effect size: 0.5
4.1.2 Social cognition
Comparing mean scores on the RMET, independent samples t-test showed no significant
difference between FBD and NBD (p < .05, table 4, page 19). After transforming the mean
RMET scores into norm scores, both groups scored within the normal range, indicating no problems recognizing emotions based on pictures of eyes. Faux pas scores in FBD patients also did not significantly differ from NBD patients, U = 355.0, z = -1.88, ns, r = .24. Although there
may be a trend in the data (p < .10), suggesting that FBD patients perform slightly worse on the
Faux pas test than NBD patients, this difference was not clinically relevant (r < .30).
Furthermore, FBD and NBD patients did not significantly differ in mean performance on the memory questions (p > .05), indicating that any difference in performance on this test is not due
to differences in remembering the stories. In summary, these results suggest that FBD patients in the chronic phase after ABI do not have more problems recognizing emotions based on pictures of eyes or recognizing and understanding faux pas stories than NBD patients. Therefore, these tests may lack sensitivity and specificity to the changes following structural frontal brain damage.
Table 4. Social cognition
FBD NBD
Independent samples t-test Mean (SD) Mean (SD) t df p r
RMET 22.48 (4.92) 25.04 (3.51) -1.58 61 0.119 0.20
Mann-Whitney tests Mean rank Mean rank U Z p
Faux pas
Faux pas memory 28.14 31.20 36.82 33.00 355.0 462.0 -1.88 -1.28 0.060 0.202 0.240.16
*Significant level set at p < .05, clinical relevance at p < 0.006 (Bonferroni correction)
*Small effect size: 0.1, medium effect size: 0.3, large effect size: 0.5
20
Contrary to predictions, independent samples t-tests showed no significant and clinically relevant
differences in mean subscale and total FrSBe standard scores (p > .05, r <.30, table 5 page 20).
These results indicate that SO’s do not rapport more behavioural problems for FBD patients in the chronic phase after ABI than for NBD patients on the FrSBe. This behavioural questionnaire may therefore lack specificity to the changes following FBD. According to the norms, mean scores of apathy, behavioural executive dysfunction, and total behavioural dysfunction scores are elevated at a clinically significant level for both groups. Although mean scores on disinhibition did not significantly differ, a trend was seen, with the average mean score of the FBD group (clinical significance range) appearing to be slightly higher than the average mean score of the NBD group (borderline of significance range).
Table 5. Behaviour FBD NBD Mean* (SD) Mean* (SD) t df p r FrSBe Apathy FrSBe Disinhibition FrSBe Executive dysfunction FrSBe Total score
72.69 (18.90) 66.97 (19.34) 69.86 (18.24) 73.51 (18.66) 77.36 (18.85) 61.46 (17.06) 71.71(16.52) 72.29 (16.75) -0.98 1.18 -0.42 0.27 61 61 61 61 0.333 0.242 0.667 0.787 0.12 0.15 0.05 0.04
*FrSBe standard t-scores
Independent samples t-tests
*Significance level set at p < .05; clinical relevance level set at p < .006 (Bonferroni correction)
*Small effect size: 0.1, medium effect size: 0.3, large effect size: 0.5
4.2 The relationship between cognition and behaviour
4.2.1 Behaviour and cognitive executive functioning
For both groups the regression models, using cognitive executive dysfunctions as the independent variables and Apathy, Executive dysfunction and total behavioural dysfunction as dependent variables, were not significant, p > .05; p > .10 (table 6 page 20). For the FBD group,
the regression model using cognitive executive dysfunction as the independent variable and Disinhibition as the dependent variable, was significant F (4, 30) = 3.10, p = .01. Although this
result indicates that the regression model - Stroop, b = .30, t = 1.83, p = .077, 95% CI [-1.17 -
21.79], TMT, b = 0.29, t = 1.78, p = .085, 95% CI [-0.86 – 12.66], WCST, b = -.20, t = -1.37, p =
.180, 95% CI [-0.81 – 0.16], and DAT, b = -0.12, t = -0.78, p = 0.443, 95% CI [-0.79 – 0.36] -
may be useful for predicting behavioural disinhibition, this prediction is weak: only 26.1% of the differences in behavioural disinhibition can be predicted based on performance on cognitive executive tests (R2 = .261). For the NBD group the regression model using cognitive executive dysfunction as the independent variables and Disinhibition as the dependent variable, was not
21
significant F(4, 22) = 0.45, p >.05; p >.10. In summary, these results indicate a dissociation
between cognition and behaviour in FBD patients in the chronic phase after ABI, and in NBD patients.
Table 6. Regression models cognitive executive functioning and behaviour for both patient groups
Apathy Disinhibition EF Total
FBD1 NBD2 FBD NBD FBD NBD FBD NBD Const. Stroop TMT WCST DAT R2 F 78.27 -6.74 1.52 0.33 -0.42 0.12 0.995 117.95 5.48 -12.20 -0.07 0.63 0.18 1.24 56.16* 10.31 5.90 -0.33 -0.22 0.26 3.99 84.30 4.07 -7.87 -0.24 -0.11 0.08 0.45 50.69 2.87 5.70 0.18 -0.25 0.19 1.72 96.08 8.75 -11.35 -0.01 -0.50 0.26 1.94 57.37 1.49 7.33 0.10 -0.28 0.23 2.25 103.53 7.03 -12.26 -0.03 -0.52 0.22 1.51 1 N = 35 2 N = 28 * p < .05 ** p < .01 *** p < .001.
4.2.2 Behaviour and social cognition
For both patient groups, the regression models using social cognitive measures as the independent variables and Apathy, Disinhibition Executive dysfunction and total behavioural dysfunction as dependent variables, were not significant (p > .05; p > .10, table 7 page 21). These
regression models are thus not useful for predicting behavioural disturbances related to frontal lobe functioning in both patient groups. Furthermore, these results indicate a dissociation between social cognition and behaviour in NBD and FBD patients in the chronic phase after ABI and patients without structural brain damage.
Table 7. Regression models social cognition and behaviour for both patient groups
Apathy Disinhibition EF Total
FBD1 NBD2 FBD NBD FBD NBD FBD NBD Constant Fauxpas RMET R2 F 78.48 -0.07 -0.21 0.00 0.06 105.41 -0.25 -0.93 0.06 0.75 73.91 -0.48 0.07 0.02 0.32 114.94 1.04 -1.31 0.17 2.57 74.76 -0.60 0.25 0.04 0.62 96.72 -0.28 -0.78 0.05 0.71 70.30 -0.46 0.50 0.03 0.55 107.87 -0.45 -1.07 0.10 1.38 1 N = 35 2 N = 28 * p < .05 ** p < .01 *** p < .001. 5. Discussion
22
This study researched the sensitivity and specificity of conventional neuropsychological measures of cognitive executive functions, social cognition and the behavioural questionnaire Frontal Systems Behaviour Scale (FrSBe; Grace & Malloy, 2001). In addition, this study researched the relationship between cognition and behaviour in FBD patients in the chronic phase after ABI. To our knowledge, this is the first study that directly related FrSBe’s proposed behavioural frontal syndromes to neuropsychological performance in a chronic outpatient group with objectified and well-defined ABI.
5.1 Conventional neuropsychological measures of cognitive executive functions
Validating predictions based on previous research (e.g. Slachevsky et al., 2006; Ardila, 2008; Namiki et al., 2008; Knutson et al., 2015; Alvarez & Emory, 2006; Caracuel, et al. 2008), this study found no significant and / or clinically relevant differences in performance on conventional tests of cognitive executive functions (i.e. TMT, WCST perseverative errors, Stroop and DAT) between the study groups. Although FBD patients performed significantly worse on a measure of verbal fluency, this result was not clinically relevant. Considering limited power due to small sample size, this data trend requires further research. According to norm tables, both patient groups scored on average within the normal range on all the tests, indicating no disturbances in cognitive executive functioning. Together, these results strongly suggest a lack of sensitivity and specificity of conventional measures of cognitive executive functions often used in clinical practice; stress the importance of developing and validating new and more ecologically valid neuropsychological methods for different groups of FBD patients in the chronic phase after ABI; e.g. ‘reversal learning paradigms’ (O’Doherty, Kringelbach, Rolls, Hornak, & Andrews, 2001; Berlin et al., 2004; Jonker, Jonker, Scheltens, & Scherder, 2015), methods using virtual / augmented reality, and different tests for ‘hot’ and ‘cold’ cognition (respectively sensitivity to reward and punishment and logical and rational thinking processes).
5.2 Social cognition
Contrary to predictions based on previous studies (e.g. Mitchell, Avny, & Blair, 2006; Namiki et al., 2008; Hornak et al., 1996; Blair & Cipollloti, 2000), this study showed no discriminative ability of social cognitive tests (Faux pas and RMET), indicating that FBD patients in the chronic phase after ABI do not have more problems recognizing emotions based on pictures of eyes and recognizing and understanding faux pas stories than NBD patients. Nevertheless, FBD patients seem to perform slightly worse on the Faux pas test than NBD patients, supported by the trend in the data. This however, needs further investigation in a larger and more homogeneous sample.
23
In the current study, the scarcity of patients with damage restricted to particular areas within the frontal lobes (e.g. DLPFC, vMPFC, ACC, and OFC) complicated subdividing the groups based on damage restricted to these particular areas. This heterogeneity can be seen as a major limitation of the current study. Previous studies have suggested that social cognitive problems are specifically related to damage to inferior medial frontal circuits (Blair, 2004). Furthermore, regarding the multidimensionality of social cognition, research has shown the importance of more specifically dividing these circuits and inferior frontal areas into the ventromedial PFC (vmPFC) and the dorsomedial PFC (dmPFC). These regions have been related to different aspects of social cognition, respectively ‘affective Theory of Mind (ToM)’ and ‘cognitive ToM’ (Elliot, Dolan, & Frith, 2000; Leopold et al., 2012; Shamay-Tsoory, Tomer, Berger, Godsher, Aharon-Peretz, 2005). Affective ToM can be defined by the ability to ascribe feelings or emotional states to others, whereas cognitive ToM can be described as a cognitive understanding of the thoughts, beliefs and intention of others (Premack & Woodruff, 1978). It has been suggested that patients with vmPFC lesions have problems representing another person's’ emotional mental state (affective ToM), but no problems representing cognitive mental states (cognitive ToM; Shamay-Tsoory, et al., 2003, 2005, 2006). Although these studies strongly suggest cerebral localization of functions, it is important to keep in mind that location-function relationships are defined by interconnected neural circuits, implying that dysfunction in one region may affect functions in other regions. This illustrates the difficulty of explaining potential unique contributions of the prefrontal cortex to cognition and behaviour and may be partially responsible for the inconclusiveness of research regarding the specific role of (areas within) the PFC. Future research should more intensively incorporate theoretical, neurophysiological, and advanced neuro-anatomical approaches, such as voxel-based morphometry (VBM; Kimberg, Coslett, & Schwartz, 2007).
5.3 Frontal Systems Behaviour Scale (FrSBe)
While previous studies show that the FrSBe can reliably distinguish FBD patients from normal controls and addicted patients (Caracuel, et al. 2008), the results of the current study suggest that the FrSBe family rating version cannot distinguish FBD patients in the chronic phase after ABI from NBD patients. Notwithstanding a lack of significant differences between the groups, which may be partly due to previously mentioned factors (heterogeneity of the sample, limited power and / or limited variance in behavioural reporting by significant others), both patient groups had - on average - clinically significant high scores on subscale and total scores of the FrSBe. These results strongly suggest a lack of specificity of this questionnaire in the population included in this
24
study (i.e. a neuropsychiatric population) and stress the importance of developing more specific measures of behavioural dysfunction following FBD. Future research should focus on accurate conceptualization and measurement of frontal behavioural disturbances to determine their prevalence and impact on activities in daily life.
5.4 Dissociation between cognition and behaviour
In line with previous research on patients with inferior FBD (e.g. Slachevsky et al., 2006; Namiki et al., 2008; Knutson et al., 2015; Funayama et al., 2010), but contrary to traditional neuropsychological assumptions, the results of the current study suggest a dissociation between cognition and behaviour, by showing no relationship between performance on conventional neuropsychological tests of cognitive executive functioning and behavioural disturbances (as measured with the FrSBe). Contrary to predictions, no relationship was found between disinhibition and social cognition, suggesting that the proposed relationship between disinhibition and social cognition may only be found in a population with damage restricted to the inferior frontal lobes / OFC (e.g. Namiki et al. 2008; Slachevsky et al., 2006; Knutson et al., 2015; Hornberger, Geng, & Hodges, 2011). Although these studies provide evidence for a mediating role of a complex neural network involving frontal, subcortical, and mesolimbic structures in social behaviour, it remains unclear how each specific brain region contributes to different behavioural abnormalities and what potential mediating factors are (Zamboni, Huey, Krueger, Nichelli, & Grafman, 2007).
5.5 Limitations
As mentioned before, an important limitation of the current study concerns the heterogeneity of the FBD sample, which may have influenced the results. This complicates drawing scientific conclusions about the aetiology of disturbances, although it may better resemble clinical practice. Since the behavioural syndromes measured by the FrSBe are – in theory – specifically related to different frontal-subcortical neural circuits (e.g. Cummings, 1993, Mega & Cummings, 2007), possible differences in behaviour between subgroups of FBD patients cannot be ruled out and need to be further studied in a larger sample.
Regarding the NBD group, the conventional neuroanatomical approach used in this study can also be seen as a limitation. Although these patients did not have structural brain damage, other potential cerebral abnormalities, such as functional disturbances due to mild traumatic brain injury or whiplash, cannot be ruled out as possible mechanisms underlying their chronic unexplained complaints and behavioural changes (Aubry, Cantu, Dvorak, Graf-Baumann,
25
Johnston, Kelly, et al., 2002; Mooney, Speed, & Sheppard, 2005). Future studies should therefore rely on specialized and functional neuroimaging techniques, such as fMRI (McAllister, Sparling, Flashman, & Saykin, 2001) and include a healthy control group or a patient group with damage restricted to areas outside of the frontal cortex.
Although the results of the current study strongly suggest that the measured constructs of the FrSBe are at least partly independent of cognitive dysfunction, potential mediating factors were not accounted for. These include, but are not limited to: personality, coping or depression (e.g. Shiehser et al., 2011) or different aspects of behavioural constructs (e.g. cognitive and emotional aspects), which have been suggested to be differentially related to certain cognitive aspects (e.g. Andersson & Bergedalen, 2002). Future studies should more specifically define these different aspects of behaviour and cognition and control for potential mediating factors.
Another limitation of this study concerns the cross-sectional design, with an average time since lesion of 12.7 years, which makes it impossible to draw conclusions about individual neuropsychiatric changes over time and the possible impact of caregiver’s distress (Kreutzer, Gervasio, & Camplair, 1994). Although we excluded patients with a history of substance abuse, neurological diseases, developmental disorders and most psychiatric diseases known to affect the frontal lobes, and controlled for depression and educational level, we did not take into account objectified information about severity of brain damage and premorbid cognitive, emotional and behavioural functioning, such as intelligence, personality and coping strategies of both patients and caregivers, which complicates the interpretation of post-lesion performance and behaviour. For example, previous research has shown that passive emotion focused coping styles in the chronic phase after ABI are related to subjective complaints and worse psychosocial outcome (Wolters, Stapert, Brands, & van Heugten, 2011). Whereas the FrSBe has previously been shown to have good construct validity; demonstrating significant pre- and post-injury behavioural differences in a group of patients with Multiple Sclerosis (MS; Chiaravalloti & DeLuca, 2003), these results have – to our knowledge – not been replicated in a chronic outpatient group with ABI.
Finally, because of the small sample sizes we did not delete outliers - and log transformations did not solve the normality problems - which may have complicated the statistical process. Otherwise, the patients included in this study can be considered representative for the great variability in clinical practice; unreliable cases were excluded from the study. Moreover, deciding whether or not to include or delete outliers can contribute to p-hacking, leading to bias in science (Head, Holman, Lanfear, Kahn, & Jennions, 2015).
26
5.6 Conclusion
Despite of the fact that a growing number of studies, including this study, have shown that conventional neuropsychological measures of ‘frontal lobe functioning’ lack sensitivity and specificity in FBD patients, these tests are still used in clinical practice and interpreted in conventional neuropsychological terms, implying causal relationships between cognitive dysfunction and behavioural disturbances. Especially in the absence of imaging material, (some) diagnosis of FBD may be missed, despite evident clinical symptoms. Observed social-behavioural symptoms may be interpreted in psychiatric / malingering terms. This may not only have major implications for treatment choice and thereby treatment efficacy, but may also have major implications in judicial situations. The results of this study emphasize the importance of developing and validating new and more ecologically valid neuropsychological methods, incorporating behavioural, cognitive, and emotional approaches, personality factors, coping styles, and advanced neuroanatomical- and neurophysiological approaches. Identification of relationships between lesion location and behavioural changes can help clinicians better predict behavioural risks and treatment planning.
Critical note
Since the neuropsychological tests and questionnaires included in this study do not clarify any underlying mechanisms of possible neurocognitive and behavioural consequences of frontal brain damage – and thereby do not provide any answers regarding the aetiology of the observations – all the results in this study should be interpreted in light of the theories and thoughts behind the investigated material. This implies that one should be careful generalizing any results in terms of behaviour and cognition out of context of the investigated material. In order to better understand the underlying mechanisms of behavioural and cognitive deficits, more advanced neuroanatomical- and neurophysiological approaches are recommended (for example VBM and fMRI).
27 6. References
Ardila, A. (2008). On the evolutionary origins of executive functions. Brain and Cognition, 68,
92-99.
Aubry, M., Cantu, R., Dvorak, J., Graf-Baumann, T., Johnston, K., Kelly, J., et al. (2002).
Summary and agreement statement of the first international conference on concussion in sport, Vienna 2001. British Journal of Sports Medicine, 36(1), 6-10.
Baron-Cohen, S., Wheelwright, S., Hill, J., Raste, Y., & Plumb, J. (2001). The “reading the mind in the eyes test revised version: a study with normal adults, and adults with Asperger syndrome or high functioning autism. Journal of Child Psychology and Psychiatry, 42(2),
241-251.
Barrash, J., Asp, E., Markon, K., Manzel, K., Anderson, S.W., & Tranel, D. (2011). Dimensions of personality disturbance after focal brain damage: investigation with the iowa scales of personality change. Journal of Clinical and Experimental Neuropsychology, 33(8), 833-852.
Berlin, H.A., Rolls, E.T., & Kischka, U. (2004). Impulsivity, time perception, emotion and reinforcement sensitivity in patients with orbitofrontal cortex lesions. Brain, 127(5),
28
Blair, R.J. (2004). The roles of orbital frontal cortex in modulation of antisocial behaviour. Brain and Cognition, 55, 198-208.
Bonelli, R.M., & Cummings, J.L. (2007). Frontosubcortical circuitry and behaviour. Dialogues in Clinical Neuroscience, 9(2), 141-151.
Carvalho, J.O., Ready, R.E., Malloy, P., & Grace, J. (2013). Confirmatory factor analysis of the frontal systems behaviour scale (FrSBe). Assessment, 20(5), 632-641.
Chiaravalloti, N.D., & DeLuca, J. (2003). Assessing the behavioural consequences of multiple sclerosis: An application of the frontal systems behaviour scale (FrSBe). Cognitive & Behavioural Neurology, 16(1), 54-67.
Elliot, R., Dolan, R.J., & Frith, C.D. (2000). Dissociable functions in the medial and lateral orbitofrontal cortex: Evidence from human neuroimaging studies. Cerebral Cortex, 10(3),
308-317.
Eslinger P.J., & Damasio A.R. (1985). Severe disturbance of higher cognition after bilateral frontal lobe ablation: patient EVR. Neurology, 35: 1731–1741.
Fernández-Abascal, E.G., Cabello, R., Fernández-Berrocal, P., & Baron-Cohen, S. (2013). Test-retest reliability of the ‘reading the mind with the eyes’ test: a one-year follow-up study.
Molecular Autism, 4(33).
Funayama, M., Mimura, M., Koshibe, Y., & Kato, Y. (2010). Squalor syndrome after focal orbitofrontal damage. Cognitive and Behavioral Neurology, 23(2), 135-139.
Gregory, C., Lough, S., Stone, V., Erzinclioglu, S., Martin, L., Baron-Cohen, S., & Hodges, J.R. (2002). Theory of mind in patients with frontal variant frontotemporal dementia and alzheimer’s disease: theoretical and practical implications. Brain, 125, 752-764.
Head, M.L., Holman, L., Lanfear, R., Kahn, A.T., & Jennions, M.D. (2015). The extend and consequences of p-hacking in science. PLOS Biology [Internet], 13(3):e1002106.
Hornberger, M., Geng, J., & Hodges, J.R. (2011). Convergent grey and white matter evidence of orbitofrontal cortex changes related to disinhibition in behavioural variant frontotemporal dementia. Brain: a Journal of Neurology, 134, 2502-2512.
Hornak, J., Rolls, E.T., & Wade, D. (1996). Face and voice expression identification in patients with emotional and behavioural changes following ventral frontal lobe
damage. Neuropsychologia, 34, 247–61.
Jonker, F.A., Jonker, C., Scheltens, P., & Scherder, E.J.A. (2015). The role of the orbitofrontal cortex in cognition and behavior. Reviews in the Neurosciences, 26(1), 1-11.
Kan, C.C. (2007). Toepassing van de sociale interpretative test en de lees-de-ogen-test bij volwassenen met een autismespectrumstoornis. Tijdschrift voor Psychiatrie, 49, 5124.
29
Kimberg, D.Y., Coslett, H.B., & Schwartz, M.F. (2007). Power in voxel-based lesion-symptom mapping. Journal of Cognitive Neuroscience, 19(7), 1067-1080.
Kok, J.S., & van Dam, B. (2012). Trail Making Test A+B. URL: http://www.normgroepen.nl/website/page18.php
Knutson, K.M., Dal Monte, O., Schintu, S., Wasserman, E.M., Raymont, V., Grafman, J., & Krueger, F. (2015). Areas of brain damage underlying increased reports of behavioral disinhibition. The Journal of Neuropsychiatry and Clinical Neurosciences, 27, 193-198.
Kreutzer, J.S., Gervasio, A.H., & Camplair, P.S. (1994). Patient correlates of caregivers’ distress and family functioning after traumatic brain injury. Brain Injury, 8(3), 211-230.
Krueger, C.E., Laluz, V., Rosen, H.J., & Neuhaus, B., Miller, B.L., & Kramer, J.H. (2011). Double dissociation in the anotamy of socioemotional disinhibition and executive functioning in dementia. Neuropsychology, 25(2), 249-259.
MacPherson, S.E., Phillips, L.H., Della Sala, S. (2002). Age, executive function and social decision making: A dorsolateral prefrontal theory of cognitive aging. Psychology and Aging, 17(4),
598- 609.
McAllister, T.W., Sparling, M.B., Flashman, L.A., & Saykin, A.J. (2001). Neuroimaging findings in mild traumatic brain injury. Journal of Clinical and Experimental Neuropsychology; 23, 775–791.
Mitchell, R.L., & Phillips, L.H. (2007). The psychological, neurochemical and functional neuroanatomical mediators of the effects of positive and negative mood on executive functions. Neuropsychologia, 45, 617-629.
Mooney, G., Speed, J., & Sheppard, S. (2005). Factors related to recovery after mild traumatic brain injury. Brain Injury, 19(12), 975-987.
Namiki, C., Yamada, M., Yoshida,H., Hanakawa,T., Fukuyama, H., & Murai, T. (2008). Small orbitofrontal traumatic lesions detected by high resolution MRI in a patient with major behavioural changes. Neurocase: The neural basis of cognition, 14(1), 474-479.
O’Doherty, J., Kringelbach, M.L., Rolls, E.T., Hornak, J., & Andrews, C. (2001). Abstract reward and punishment representations in the human orbitofrontal cortex. Nature Neuroscience, 4,
95-102.
Ready, R.E., Stierman, L., & Paulsen, J.S. (2010). Ecological validity of neuropsychological and personality measures of executive functions. The Clinical Neuropsychologist, 15(3), 314-323.
Rolls, E.T. (1996). The orbitofrontal cortex. Philosophical Transactions of the Royal Society of London Series B- Biological Sciences, 351(1346), 1433–1444.
