1
Identification of cognitive biomarkers in presymptomatic
familial frontotemporal dementia – a longitudinal follow-up
study
Master thesis
Lauren van Asseldonk
Student number: 11147024
Master thesis Clinical Neuropsychology
University of Amsterdam
Date: June 16
th, 2017
Daily supervisor Erasmus Medical Center (Erasmus MC): Ms. Drs. L. C. Jiskoot, Department of Neurology of Erasmus MC and Department of Radiology of the Leiden University Medical Center (LUMC).
Supervisor University of Amsterdam: Dr. F. van Opstal, faculty of social and behavioural sciences, department of brain & cognition.
2
Abstract
Objective: In this longitudinal study we assessed neuropsychological functioning in presymptomatic
familial FTD during four years in order to investigate presymptomatic cognitive decline and to identify neuropsychological domains and tests that could differentiate between converters and
non-converters before and at symptom onset.
Methods: The current study is part of the Frontotemporal Dementia Risk Cohort (FTD-RisC) and
assessed cognitive functioning in six neuropsychological domains through a baseline assessment and two follow-up assessments after two and four years in MAPT mutation carriers (n=15), GRN mutation carriers (n=31) and healthy controls (n=39). Eight mutation carriers converted to the disease phase during the study time window. For our longitudinal objective, we analysed cognitive decline over time through multilevel regression analyses. Secondly, we performed cross-sectional ROC-analyses in order to identify neuropsychological tests that could differentiate between symptomatic and
presymptomatic mutation carriers. At last, we performed logistic regressions in order to find an optimal combination of tests that could be of use in predicting or diagnosing FTD.
Results: Symptomatic participants deteriorated over time within the domains of social cognition,
attention and mental processing speed, executive functioning, memory, and language. Furthermore, we found gene-specific decline patterns in MAPT and GRN converters, with MAPT converters showing a decline over time within the domains of social cognition, attention and mental processing speed, executive function, memory, and language, while GRN converters deteriorated within the domains of attention and mental processing speed and executive functioning over time. Explorative analyses showed that categorical fluency was able to differentiate between symptomatic and presymptomatic mutation carriers at symptom onset, but not as early as four or two years before symptom onset.
Conclusions: The current study showed subtle presymptomatic cognitive decline in genetic FTD and
gene-specific patterns of decline in MAPT and GRN mutation carriers. We found neuropsychological measures not to be predictive for symptom onset, but to be of considerable use in classifying and diagnosing FTD converters at symptom onset. These findings can contribute substantially to the recognition of FTD cases in the early disease stage, which confirms the value of neuropsychological assessment in facilitating early FTD diagnostics.
3
Keywords: frontotemporal dementia; presymptomatic cognitive decline; cognitive biomarkers;
4
IntroductionFrontotemporal dementia (FTD) is a neurodegenerative disease spectrum in which
behavioural changes and/or impairment in language functions are the presenting symptoms (Warren, Rohrer, & Rossor, 2013). The prevalence of FTD is 15 in every 100.000 individuals, with symptoms emerging at a relatively early age of 45-65 years old on average (Neary, Snowden, & Mann, 2005; Ratnavalli, Brayne, Dawson, & Hodges, 2002). The term
‘frontotemporal’ refers directly to brain parts that are considered to be most involved in the disease process – the frontal and temporal lobes (Warren, Rohrer, & Rossor, 2013).
The identified neuroimaging patterns are inherent to the deterioration of neuropsychological functions found by clinical studies (Rascovsky et al., 2011; Seelaar, Rohrer, Pijnenburg, Fox, & van Swieten, 2011).The clinical problems the patient experiences depend on the form in which FTD manifests. The two main forms are behavioural variant FTD (bvFTD) and primary progressive aphasia (PPA). BvFTD primarily presents with
behavioural changes, which are associated with frontal atrophy (Warren, Rohrer, & Rossor, 2013). Language functions are primarily compromised in PPA, which corresponds with temporal atrophy (Warren, Rohrer, & Rossor, 2013). However, additional cognitive domains can decline in later disease stages (Seelaar, Rohrer, Pijnenburg, Fox, & van Swieten, 2010; Warren, Rohrer, & Rossor, 2013). Impairment in social cognition, executive functions, and language abilities are generally most apparent in patients with early FTD, while memory and visuospatial functions are considered to remain relatively intact (Rascovsky et al., 2011; Seelaar, Rohrer, Pijnenburg, Fox, & van Swieten, 2010). FTD diagnosis is challenging due to the insidious onset, overlap with other forms of dementia and psychiatric phenomena. As a result, FTD often remains unrecognized in the early disease phase, causing a substantial diagnostic delay (Mathuranath et al., 2000; Ratnavalli, Brayne, Dawson, & Hodges, 2002).
5
A genetic cause in the form of an autosomal dominant gene mutation is discernible in up to 40% of FTD cases (Seelaar, Rohrer, Pijnenburg, Fox, & van Swieten, 2010; Warren, Rohrer, & Rossor, 2013). An autosomal dominant inheritance pattern implies that children of mutation carriers have a 50% chance of developing FTD themselves (Chow, Miller, Hayashi, & Geschwind, 1999). Almost 50% of the gene mutations are present in
microtubule-associated protein tau (MAPT) and progranulin (GRN) genes (Seelaar, Rohrer, Pijnenburg, Fox, & van Swieten, 2010). A later discovered gene mutation is in chromosome 9 open reading frame 72 (C9orf72). BvFTD is the most frequently manifested clinical form in both MAPT and GRN gene mutations (Pickering-Brown et al., 2008). MAPT mutation carriers frequently present with language impairment — specifically semantic deterioration—, behavioural disinhibition, and obsessive-compulsive behaviour. In GRN mutation carriers, impairment in language and episodic memory are more characteristic. In addition, delusions and hallucinations are psychiatric features that are frequently associated with mutations in GRN genes (Seelaar, Rohrer, Pijnenburg, Fox, & van Swieten, 2010).
A development in this field of research is studying the presymptomatic phase in FTD. Studying the presymptomatic phase in FTD is highly relevant. First of all, identification of biomarkers is essential for tracking disease onset and progression, and biomarkers can be of considerable use as clinical endpoints in clinical trials. Secondly, studying the
presymptomatic phase allows us to detect changes in brain functionality in a phase in which brain damage is still limited. The ideal starting point for future medication trials would be the moment the earliest pathological processes begin. Up until this day no cure for FTD has been proved effective, but more knowledge about disease processes in FTD may offer starting points for development of future therapies.
Different areas of research have confirmed changes in brain and cognition in the years before symptom onset, which offers starting points for the development of biomarkers in different disciplines. Firstly, research groups have shown that significant patterns of
6
decline are present in multiple neuropsychological domains in presymptomatic mutation carriers (Jiskoot et al., 2016; Rohrer et al., 2015). Furthermore, structural and functional brain anomalies were identified by neuro-imaging studies in presymptomatic mutation carriers, and decreased cerebral blood flow (CBF) was correlated with deteriorated
performances on specific neuropsychological tests (Dopper et al., 2013; Dopper et al., 2016). Lastly, progranulin protein levels in cerebrospinal fluid are lower in presymptomatic GRN mutation carriers than in healthy controls (Meeter et al., 2016).
Elaborating on cognitive decline in the presymptomatic phase, Jiskoot et al. (2016) found that performances on language and social cognition tests already start to decline years before symptom onset. Surprisingly, Jiskoot et al. (2016) also found significant
deterioration in visuoconstruction and memory functions, which is contradictory to previous findings and formal diagnostic criteria (Rascovsky et al., 2011; Seelaar, Rohrer, Pijnenburg, Fox, & van Swieten, 2010). Jiskoot et al. (2016) explained these findings in the light of a paper by Irwin et al. (2016), which stated that visuospatial dysfunctions were also found in later disease phases in Pick disease—another other FTD-tauopathy—, and through a paper by and Collette et al. (2010), which implied that memory dysfunctions might be the result of compromised executive function— for instance impaired retrieval strategies. The study by Rohrer et al. (2015) showed significant decline on naming tasks and pointed towards deterioration in executive function. From a gene-specific view, Jiskoot et al. (2016) found MAPT mutation carriers to present with lower language and mental processing speed performances, and to deteriorate faster in social cognition than GRN mutation carriers. As for GRN mutation carriers, Barandiaran et al. (2012) showed that GRN mutation carriers show worse performances on tests for attention, mental flexibility, and naming than healthy controls. In order to detect patterns of decline years before presentation of the first
symptoms, Jiskoot et al. (2016) made individual estimations of the years before symptom onset. This estimation is a common approach and was based on the average age at which
7
symptoms manifest in the family, which is often related to the individual age of onset in autosomal dominant FTD (Rohrer et al., 2015). Jiskoot et al. (2016) found that language, social cognition, and memory domains showed deterioration five to seven years before expected symptom onset in MAPT mutation carriers. In GRN mutation carriers, memory was found to deteriorate eight years in advance, and five years before symptom onset
performances on a particular test for social cognition (Happé non-ToM, Happé et al., 1999) started to decline. Overall, more research with longer follow-up is needed in order to confirm a consistent neuropsychological decline pattern in presymptomatic FTD mutation carriers.
Up until this day, few studies have focused on neuropsychological functioning in the presymptomatic phase. As little as two studies found that there are signs of cognitive decline in the presymptomatic phase (Jiskoot et al., 2016; Rohrer et al., 2015), but both studies have substantial limitations. Rohrer et al. (2015) used a cross-sectional design, and whereas the study by Jiskoot et al. (2016) is truly longitudinal, estimated ages of onset were used. Estimated ages of onset are commonly used, but are relatively less accurate measures. The current study elaborates on previous research by using a longitudinal design with two follow-up assessments and by using actual ages of onset.
The current study is part of the Frontotemporal Dementia Risk Cohort (FTD-RisC) (Dopper et al., 2014; Meeter et al., 2016; Jiskoot et al., 2016). FTD-RisC is the first research group to study e.g. neuropsychological functioning in presymptomatic mutation carriers of autosomal dominant FTD to this extent. The current study elaborates on previous research by Jiskoot et al. (2016) and is similar regarding research design. Longitudinal measures in presymptomatic at-risk individuals are performed through a baseline assessment and two follow-up assessments after two and four years. The main objective of the current study is to contribute to previous research by trying to identify presymptomatic, gene-specific changes in neuropsychological test profiles in MAPT and GRN mutation carriers. In addition, we aim
8
to identify neuropsychological tests that could potentially differentiate between
symptomatic and non-symptomatic mutation carriers four years before symptom onset, two years before symptom onset and at the moment of conversion. Based on literature, we expect to see a pattern of presymptomatic neuropsychological decline over time in mutation carriers compared to healthy controls, particularly within the social cognition, executive function, and language domains. More specifically, we expect MAPT mutation carriers to show the earliest neuropsychological decline within the domains of social cognition, mental processing speed, and language. For GRN mutation carriers we expect to see patterns of neuropsychological decline on measures of attention, mental flexibility and naming. For the second, more explorative objective of the current study to identify neuropsychological tests that are particularly sensitive to FTD development, we expect tests within the above
mentioned domains to show potential of contributing to the tracking and onset of disease processes in FTD.
Methods Participants
At-risk, native-Dutch individuals from genetic GRN or MAPT FTD families were asked to participate in the study. For inclusion, participants needed to be at least 18 years old and they were required to be asymptomatic at baseline, regarding the presymptomatic nature of the study. Participants were considered presymptomatic when not all formal FTD criteria were met, when no cognitive disorders could be established by psychometric assessment (i.e. scores equal to or lower than 2 SD’s below normative means) and when there were no signs of behavioural changes or impairment in cognitive functioning. Participants were excluded from baseline- and follow-up measures if they met the criteria for or had a history of neurologic disorders, major psychiatric disorders or drug abuse. Participants were tested three times, once every two years. All clinical researchers were undisclosed to information
9
about the participants’ mutation carrierships. A total of 85 participants were included in the analysis, of which 15 MAPT mutation carriers, 31 GRN mutation carriers, and 39
non-mutation carriers from the same families, the latter forming a healthy control group. Eight participants converted to the disease stage during the study, of which five MAPT mutation carriers and three GRN mutation carriers. The participants signed an informed consent form in advance, in which they approved of the terms of participating in the study. Two control subjects were excluded from analyses, as they performed at disorder-level on
neuropsychological assessment and were suspected to have developed a different type of dementia. Therefore, these participants were not suitable as healthy controls. If participants dropped-out due to too advanced disease processes, the cases were not excluded from analyses. These cases show the furthest deterioration and are therefore most informative about FTD disease processes.
Figure 1
10
Operationalisation
The current study is part of the FTD-RisC research initiative at the Erasmus Medical Center. Within FTD-RisC, at-risk individuals from Dutch FTD families are followed and assessed once every two years. During their two-yearly visit, participants are thoroughly screened by means of neurological and neuropsychological assessments, blood samples, MRI scans, and in some cases, skin biopsies and lumbar punctures. The emphasis of the current study is on the data from an extensive neuropsychological test battery conducted at baseline and follow-up assessments that were analysed in order to detect cognitive deterioration
between these measures. Previous studies showed cognitive decline in specific domains and tests, but since this area of research is still relatively new, additional domains were included in the current study for explorative purposes. In total, a wide range of 24 tests was included, covering six neuropsychological domains (see below).
Material1
Sample characteristics
Global cognitive functioning
The Mini Mental State Examination (MMSE) and the Frontal Assessment Battery were used as screeners for cognitive functioning and frontal lobe functioning, respectively. Both screeners were conducted by the researcher.
Psychiatric screeners
Two psychiatric screeners in the form of questionnaires were conducted. The Beck Depression Inventory (BDI) measures depressive symptoms and was completed by the participant. The Neuropsychiatric Inventory (NPI) measures neuropsychiatric symptoms and
1
11
was filled out by a partner or relative.Dependent variables
Neuropsychological functioning
Neuropsychological functioning was assessed by a range of neuropsychological tests. All neuropsychological tests were classified into six different domains. The social cognition domain comprised the Happé Theory of Mind, Happé non-Theory of Mind and Ekman faces. The executive function domain entailed Trail Making Test-B (TMT-B), Stroop card III,
Modified Wisconsin Card Sorting Test (M-WCST) concepts, letter fluency, and Wechsler Adult Intelligence Scale (WAIS) similarities. For attention and mental processing speed, Trail Making Test-A (TMT-A), Stroop card I, Stroop card II and Letter Digit Substitution Test (LDST) were used. WAIS Block Design and Clock Drawing measured Visuoconstruction. The memory domain comprised the Visual Association Test (VAT), WAIS digit span, and Rey Auditory Verbal Learning Test (RAVLT). The language measures used were the Boston Naming Test (BNT), Semantic Association Test (SAT) verbal, ScreeLing-phonology and categorical fluency (animals). For further descriptions of the tests, Appendix A can be consulted at the end of the manuscript.
Data analysis
According to G*Power 3.1 (Paul, Erdfelder, Buchner, & Lang, 2009), a sample size of 65 should be sufficient to detect a medium to large effect (.5) with a power of .95 and a significance level of 0.0083 (0.05 corrected for multiple comparisons). The current study included 85 participants, which exceeds the number of participants needed to meet the power criterion.
Analyses were performed using SPSS Statistics 21.0 (IBM Corp., Armonk, NY). A total of 85 participants were included in the analyses. For all statistical tests a two-sided .0083
12
level of significance was used, since we expected decline for mutation carriers on particular domains, but learning effects could not be ruled out in advance. We calculated domain scores for each cognitive domain at each time point. The average z-score of all tests within a specific domain was considered the domain score. For non-converters – mutation carriers who did not develop the disease yet—the three time points were baseline measurement, follow-up 1 and follow-up 2. For individuals that developed FTD during the study, the converters, the time levels were classified differently (Figure 2). Because we knew between which measurements the individual converted to the disease stage (based on clinical impression, neuropsychological assessment, MRI, a.o.), their time levels were classified as four years before symptom onset, two years before symptom onset, and time of conversion. For an individual that converted between first and second follow-up measures, baseline is therefore regarded the time point of four years before symptom onset, follow-up 1 is considered the time point of two years before symptom onset, and follow-up 2 is taken as the time of conversion. Naturally, this is a rough classification, but we believe it is justified as it is not possible to determine an exact time of conversion. We made one exception to this classification format for a participant who had competed third follow-up measures and who converted between follow-up 2 and follow-up 3. For this participant, we excluded baseline measurement and included measurement at third follow-up instead.
13
Figure 2Classification of participants
Cross-sectional analyses were performed by means of one-way ANOVA’s to compare age, cognitive screeners, and psychiatric screeners between MAPT mutation carriers, GRN mutation carriers, and healthy controls. Non-parametric chi-squared tests were performed to look at differences in distributions of sex and level of education between groups.
The first research question was tested nonparametrically, as various assumptions of planned ANCOVA’s were violated. Among others, the assumptions of homogeneity of regression slopes and homogeneity of variance were partly, but not fully met. As an alternative to ANCOVA’s, multilevel regression analyses were computed to analyse
deterioration in neuropsychological test performances over time for each group. Multilevel regression analyses correct for selective missing values and do not require fulfilment of as many strict assumptions as ANCOVA’s do. In a first analysis, both gene mutations were included as fixed effects in interaction with time in order to investigate decline patterns in MAPT and GRN mutation carriers separately. This analysis was performed twice, the second
14
time excluding all symptomatic participant in order to control for their influence on the effect. In a second analysis, conversion statuses were included as fixed effects in interaction with time to distinguish between converters and non-converters. Age, sex, and level of education were included as covariates in all analyses. The dependent variables were the six neuropsychological domain scores. In order to correct for multiple comparisons, the significance level (α) was divided by the number of comparisons, in this case 6. Secondary analyses were computed for all tests belonging to domains that were found do decline over time in either type of gene mutation or conversion status for explorative purposes. In the second (explorative) analysis, the group of converters was split into MAPT and GRN mutation carriers in order to establish gene-specific decline patterns.
In pursuit of determining which tests were able to differentiate between converters and non-converters, ROC-analyses were performed. Firstly, domain scores were analysed as test variables with conversion as the outcome measure. Analyses were performed for each time point. In secondary analyses, we included all tests belonging to the domains that were found to be able to differentiate between converters and non-converters, again with conversion as the outcome measure. In order to investigate whether a combination of neuropsychological tests would create superior predictions over individual tests, we performed a forward stepwise logistic regression in a final analysis. We included all significant tests identified by the ROC-analyses as predictors, with conversion status as the outcome measure. Age, sex, and level of education were included as covariates.
Results
Table 1 presents the results of cross-sectional analyses including demographic statistics, psychiatric screeners, and depressive screeners. Age differed significantly between groups (F(2,82)=4.713, p=0.012), and Scheffé post-hoc analyses showed that MAPT mutation carriers were significantly younger than GRN mutation carriers (p=0.012). NPI scores differed
15
significantly at second follow-up measures (F(2,64)=7.558, p=0.001), with MAPT mutation carriers showing significantly more neuropsychiatric symptoms at second follow-up on the NPI than both GRN mutation carriers (p=0.007) and healthy controls (p=0.002). There were no differences in distributions of sex or level of education between the three groups. None of the groups differed on the cognitive screeners or BDI.
Table 1
Demographics, cognitive screeners and psychiatric screeners MAPT mutation carriers (n=15) GRN mutation carriers (n=31) Healthy controls (n=39) p value Sex (male) 8 (53%) 11 (35%) 17 (44%) 0.506
Age at study start 41.9 ± 10.0 52.1 ± 8.2 49.1 ± 12.2 0.012* Verhage education 5.1 ± 1.6 5.7 ± 0.9 5.2 ± 1.0 0.102 MMSE Baseline 29.6 ± 0.5 29.1 ± 1.6 29.1 ± 1.3 0.451 Follow-up 1 28.7 ± 2.2 28.9 ± 1.6 29.2 ± 1.3 0.513 Follow-up 2 28.4 ± 1.5 29.2 ± 1.4 29.2 ± 1.0 0.099 FAB Baseline - - - 0.417 Follow-up 1 17.4 ± 0.8 17.5 ± 0.9 17.4 ± 0.9 0.883 Follow-up 2 16.5 ± 1.6 17.0 ± 1.1 16.7 ± 1.7 0.639 NPI Baseline 1 4.6 ± 11.2 1.4 ± 3.4 0.13 ± 0.5 0.180 Follow-up 1 6.4 ± 20.7 0.25 ± 0.7 0.61 ± 1.2 0.095 Follow-up 2 12.3 ± 18.7 2.1 ± 6.6 0.8 ± 1.5 0.001* BDI Baseline 1 4.0 ± 6.3 3.2 ± 3.9 4.1 ± 4.5 0.693 Follow-up 1 4.5 ± 5.0 3.2 ± 4.0 3.7 ± 3.9 0.638 Follow-up 2 7.6 ± 9.5 3.0 ± 6.7 3.5 ± 4.3 0.108
Note. MMSE = Mini Mental State Examination; BDI = Beck Depression Inventory.
*Significant at a .05 level
Multilevel regression analyses showed significant deterioration over time in MAPT mutation carriers regarding performances within the domains of social cognition, memory and language (Table 2). Results of the explorative multilevel regression analyses are shown in Table 3. Within the domain of social cognition, test performances deteriorated
16
performances on VAT and RAVLT- delayed recall declined significantly over time. Within the language domain, BNT and categorical fluency scores decreased over time. Regarding GRN mutation carriers, there were no patterns of decline whatsoever, but performances on the SAT tended to improve significantly over time. As for healthy controls, test scores on Happé non-ToM, Ekman faces, RAVLT-immediate recall and RAVLT-delayed recall improved
significantly over time. After correcting for multiple comparisons, MAPT mutation carriers still showed declined performances on RAVLT-delayed recall and healthy controls still improved on RAVLT-immediate and RAVLT-delayed recall. In order to control for the
possibility that the converters were the cases driving the effects in MAPT and GRN mutation carriers, we performed the analyses without the symptomatic participants. After exclusion of converters, none of the domain scores continued to show significant decline over time. Regarding individual tests, performances on RAVLT-delayed recall still declined significantly over time in MAPT mutation carriers (=-0.032, t(112)=-2.304, p=0.023). In GRN mutation carriers, the improvement on the SAT remained significant (=0.019, t(152)=2.101, p=0.037).
Table 2
Neuropsychological test performances at baseline and change estimates over time (domain specific)
MAPT mutation carriers (n=15) GRN mutation carriers (n=31) Healthy controls (n=39)
Domain Baseline scores (SD) Change estimate p Baseline scores (SD) Change estimate p Baseline scores (SD) Change estimate p Social cognition 0.228 (0.72) -0.009 0.007* 0.332 (0.68) -0.003 0.332 0.00 (0.82) -0.000 0.878 Attention/mental processing speed 0.267 (0.58) -0.003 0.096 0.060 (0.88) -0.003 0.075 0.00 (0.81) -0.002 0.084 Executive function 0.275 (0.60) -0.005 0.065 0.221 (0.80) -0.004 0.052 0.00 (0.65) -0.001 0.505 Memory 0.089 (1.32) -0.017 <0.001* 0.117 (0.86) -0.001 0.745 0.00 (0.73) -0.000 0.848 Visuoconstruction -0.151 (0.70) -0.005 0.266 0.032 (1.05) -0.000 0.963 0.00 (0.77) -0.001 0.656 Language 0.227 (0.58) -0.010 0.002* 0.078 (0.66) -0.004 0.121 0.00 (0.63) -0.000 0.931
Note. All domain scores are z-scores
17
Table 3Neuropsychological test performances at baseline and change estimates over time (test specific)
MAPT mutation carriers (n=15) GRN mutation carriers (n=31) Healthy controls (n=39) Domain Test Baseline scores (SD) Change estimate p Baseline scores (SD) Change estimate p Baseline scores (SD) Change estimate p Social cognition Happé ToM 12.62 (3.66) -0.044 0.011 12.87 (2.91) -0.005 0.707 11.79 (3.43) 0.013 0.172 Happé non-ToM 12.38 (2.82) -0.036 0.017 13.03 (2.59) -0.012 0.331 11.69 (2.90) 0.020 0.013 Ekman faces 47.00 (5.46) -0.028 0.293 47.10 (5.47) -0.013 0.548 45.69 (6.40) 0.038 0.009 Memory VAT 11.38 (1.56) -0.012 0.019 11.53 (0.90) 0.000 0.926 11.79 (0.57) 0.001 0.740 RAVLT-imm. recall 47.46 (9.66) -0.076 0.090 46.33 (10.62) -0.015 0.686 42.64 (9.83) 0.157 <0.001* RAVLT-del. recall 9.69 (3.88) -0.048 <0.001* 9.43 (3.33) -0.000 0.983 8.44 (3.15) 0.050 <0.001* RAVLT-recognition 29.00 (2.00) -0.022 0.176 29.17 (1.18) -0.009 0.505 28.62 (2.11) 0.014 0.127 Language
Boston Naming Test 52.62 (5.30) -0.080 0.005 55.13 (3.70) 0.006 0.786 53.38 (4.50) 0.026 0.105
SAT 27.92 (1.50) -0.008 0.604 27.50 (2.01) 0.019 0.033 27.82 (1.14) -0.003 0.604
ScreeLing-phonology 23.92 (0.28) -0.005 0.190 23.75 (0.45) -0.001 0.863 23.54 (0.81) 0.001 0.733
Categorical fluency 26.46 (6.62) -0.087 0.006 23.37 (5.67) 0.021 0.424 23.87 (4.94) 0.026 0.141
Note. Happé ToM = Happé Theory of Mind; Happé non-ToM = Happé non-Theory of Mind; RAVLT = Rey Auditory Verbal Learning Test; imm. Recall =
immediate recall; del. recall: delayed recall; SAT = Semantic Association Test. *Significant on a 0.0045 level (corrected for multiple comparisons)
Table 4 shows the results of the multilevel regression analyses for the group of mutation carriers only, split into converters and non-converters. Performances on the domains of social cognition, attention and mental processing speed, executive function, memory, and language deteriorated significantly over time in converters. Results of the explorative multilevel regression analyses are shown in Table 5. Many test performances tended to decline over time in converters, even after correcting for multiple comparisons (p<0.0023). Within the social cognition domain, test performances on Happé ToM, Happé non-ToM, and Ekman faces tended to decline significantly over time in converters. Similarly, performances on Stroop card II and LDST declined within the attention and mental
processing speed domain, and for executive function we found deterioration on TMT-B, Stroop card III, M-WCST concepts, letter fluency, and WAIS similarities. VAT,
RAVLT-18
immediate and RAVLT-delayed recall were found to decline within the memory domain, and lastly, within the language domain, we found a decline in test performances on BNT and categorical fluency over time in converters. After controlling for multiple comparisons the decline remained significant on a majority of tests, except for Happé ToM, LDST, M-WCST concepts and VAT. None of the domains showed deteriorated performances over time in non-converters. Healthy controls improved on Happé non-ToM, Ekman faces, Stroop Card I and III, letter fluency, RAVLT-immediate recall and RAVLT-delayed recall. After correcting for multiple comparisons, improvement on letter fluency, immediate recall and RAVLT-delayed recall remained significant in healthy controls.
Table 4
Deterioration/improvement in neuropsychological test performances over time in converters and non-converters (domain specific)
Converters (n=8) Non-converters (n=38) Healthy controls (n=39)
Domain Baseline scores (SD) Change estimate p Baseline scores (SD) Change estimate p Baseline scores (SD) Change estimate p Social cognition 0.273 (0.90) -0.021 <0.001* 0.305 (0.67) -0.002 0.336 0.00 (0.81) -0.000 0.866 Attention/mental processing speed 0.293 (0.52) -0.012 <0.001* 0.095 (0.84) -0.001 0.321 0.00 (0.81) -0.002 0.067 Executive function 0.582 (0.33) -0.024 <0.001* 0.181 (0.775) -0.001 0.515 0.00 (0.65) -0.001 0.421 Memory -0.443 (1.84) -0.034 <0.001* 0.198 (0.81) -0.003 0.473 0.00 (0.73) -0.001 0.839 Visuoconstruction 0.213 (0.66) -0.010 0.102 -0.061 (0.99) -0.000 0.895 0.00 (0.77) -0.001 0.648 Language 0.294 (0.63) -0.019 <0.001* 0.095 (0.639) 0.002 0.408 0.00 (0.63) -0.000 0.930
Note. All domain scores are z-scores
19
Table 5Deterioration/improvement in neuropsychological test performances over time in converters and non-converters (test specific)
Converters (n=8) Non-converters (n=38) Healthy controls (n=39)
Domain Test Baseline scores (SD) Change estimate p Baseline scores (SD) Change estimate p Baseline scores (SD) Change estimate p Social cognition Happé ToM 12.67 (4.08) -0.065 0.008 12.81 (3.00) -0.012 0.380 11.79 (3.43) 0.013 0.172 Happé non-ToM 13.33 (2.50) -0.069 0.001* 12.76 (2.69) -0.012 0.267 11.69 (2.90) 0.020 0.012 Ekman faces 45.67 (5.82) -0.119 0.001* 47.30 (5.38) -0.001 0.965 45.69 (6.41) 0.038 0.007 Attention/mental processing speed
TMT-A 21.67 (6.68) 0.122 0.082 31.08 (11.77) 0.051 0.181 31.82 (15.03) -0.022 0.415 Stroop card I 44.83 (5.115) 0.065 0.087 44.38 (8.90) -0.020 0.345 47.10 (7.97) 0.039 0.010 Stroop card II 57.17 (5.88) 0.246 <0.001* 58.84 (12.90) -0.032 0.217 58.51 (10.56) 0.013 0.487 Digit span forward 9.33 (1.37) -0.019 0.168 9.27 (2.60) -0.013 0.088 8.69 (1.92) 0.001 0.868 LDST 34.83 (5.19) -0.082 0.009 33.27 (6.86) 0.004 0.809 34.54 (6.836) 0.002 0.893 Executive function
TMT-B 54.00 (25.88) 0.599 0.002* 71.22 (40.44) -0.132 0.195 67.77 (29.29) 0.052 0.470 Stroop card III 87.17 (18.50) 0.536 <0.001* 93.65 (24.75) -0.026 0.577 93.74 (22.64) -0.087 0.010 Digit span backward 7.17 (1.72) -0.027 0.072 6.51 (2.01) -0.003 0.721 6.08 (1.99) 0.008 0.188 M-WCST concepts 6.00 (0.00) -0.290 0.010 5.70 (0.81) -0.006 0.323 5.49 (0.85) 0.002 0.595 Letter fluency 39.00 (11.21) -0.254 0.001* 37.86 (12.95) -0.048 0.245 32.13 (9.85) 0.134 <0.001* WAIS similarities 29.00 (1.10) -0.154 <0.001* 25.49 (4.03) 0.004 0.775 24.77 (4.73) 0.005 0.636 Memory VAT 10.67 (2.16) -0.020 0.007 11.62 (0.83) -0.002 0.705 11.79 (0.57) 0.001 0.746 RAVLT-imm. recall 46.50 (12.69) -0.200 0.001* 46.70 (10.0) -0.009 0.797 42.64 (9.83) 0.157 <0.001* RAVLT-del. recall 8.50 (5.05) -0.063 0.001* 9.68 (3.20) -0.009 0.359 8.44 (3.15) 0.050 <0.001* RAVLT-rec. 28.17 (2.79) -0.042 0.066 29.27 (1.10) -0.009 0.461 28.62 (2.11 0.014 0.127 Language BNT 55.33 (5.72) -0.163 <0.001* 54.22 (4.15) -0.001 0.960 53.38 (4.50) 0.025 0.100 SAT 29.33 (1.37) -0.011 0.466 27.68 (1.95) 0.013 0.127 29.82 (1.36) -0.003 0.600 ScreeLing-phon. 24.00 (0.00) -0.004 0.415 23.77 (0.44) -0.002 0.551 23.54 (0.81) 0.001 0.722 Cat. fluency 26.50 (3.99) -0.200 <0.001* 23.95 (6.30) 0.014 0.546 23.87 (4.94) 0.025 0.123
Note. Happé ToM = Happé Theory of Mind; Happé non-ToM = Happé non-Theory of Mind; TMT = Trail Making Test; LDST = Letter Digit Substitution Test;
M-WCST = Modified Wisconsin Card Sorting Test; WAIS = Wechsler adult intelligence scale; VAT = Visual Association Test; RAVLT = Rey Auditory Verbal Learning Test; imm. recall = immediate recall; del. recall = delayed recall; BNT = Boston Naming Test; SAT = Semantic Association Test; phon. = phonology; Cat. fluency = categorical fluency.
*Significant on a 0.0023 level (corrected for multiple comparisons)
When we specified the analysis to type of gene mutation for explorative purposes, we saw that converters with a MAPT gene mutation were found to show a significant decline over time on performances within the domains of social cognition, attention and mental
20
processing speed, executive function, memory and language. More specifically,
performances declined on the Happé ToM (p=0.011), Stroop card I (p=0.017), Stroop card II (p<0.001), LDST (p=0.035), Stroop card III (p=0.021), WAIS similarities (p<0.001), RAVLT-immediate recall (p=0.004), RAVLT-delayed recall (p=0.030), BNT (p<0.001), and categorical fluency (p=0.001). After controlling for multiple comparisons, decline on Stroop card II, WAIS similarities, BNT, and categorical fluency remained significant. Converters with a GRN gene mutation performed more poorly over time on the domains of attention and mental processing speed and executive function. Within these domains, performances on TMT-B (p<0.001), Stroop card III (p<0.001), WCST (p=0.005), letter fluency (p=0.012) and WAIS similarities (p<0.001) deteriorated significantly over time. Decline on TMT-B, Stroop card III, and WAIS similarities remained significant after controlling for multiple comparisons. Figure 3 illustrates deterioration trends in converters and non-converters in comparison to controls on a selection of test.
21
Figure 3Deterioration trends in converters and non-converters within a selection of tests
Social Cognition Attention/Mental processing speed Executive function Memory Language Visuoconstruction
22
Results from further explorative analyses are shown in Table 6. There were no tests that were able to differentiate between converters and non-converters at four and two years before symptom onset. Various tests tended to be able to differentiate between converters and non-converters at the moment of conversion. Firstly, the domain scores of social cognition (AUC=0.826, CI=0.601-1.00, p=0.005), executive function (AUC=0.777, CI=0.568-0.985, p=0.016) and language (AUC=0.894, CI=0.795-0.993, p=0.001) tended to be able to differentiate between converters and non-converters. Within these domains, Ekman faces, Happé-ToM, Happé non-ToM (Social Cognition), Stroop card III, M-WCST concepts, letter fluency, WAIS similarities (Executive function), and categorical fluency (Language) showed potential of distinguishing between converters and non-converters. After controlling for multiple comparisons (p<0.0038), categorical fluency remained significant. Furthermore, while the majority of tests show fair AUC values, the only test with an excellent AUC value was categorical fluency (AUC=0.915, p <0.001, sensitivity=87.5%, specificity=93.9%).
23
Table 6Sensitivity and specificity of neuropsychological tests in determining FTD at disease onset
Tests AUC Confidence
Interval 95% p Optimal cut-off Sensitivity Specificity Social Cognition (z) Ekman faces 0.787 0.592-0.982 0.013 47 75.0% 78.1% Happé ToM 0.797 0.609-0.986 0.010 12 87.5% 75.8% Happé non-ToM 0.792 0.549-1.000 0.011 10 75.0% 93.9% Executive function (z)
Trail Making Test B 0.677 0.426-0.929 0.144 - - -
Stroop card III 0.790 0.604-0.975 0.012 99 75.0% 75.8%
Digit span backward 0.634 0.384-0.885 0.243 - - -
M-WCST concepts 0.732 0.502-0.963 0.044 5 62.5% 87.5% Letter fluency 0.778 0.624-0.933 0.016 34 75.0% 75.8% WAIS similarities 0.784 0.566-1.000 0.014 20 75.0% 78.8% Language (z) BNT 0.817 0.521-0.915 0.059 - - - SAT 0.585 0.367-0.803 0.459 - - - ScreeLing-phonology 0.559 0.337-0.780 0.610 - - - Categorical fluency 0.915 0.805-1.000 <0.001* 19 87.5% 93.9%
Note. (z) = z-scores; Happé ToM = Happé Theory of Mind; Happé non-ToM = Happé non-Theory of Mind; M-WCST = Modified Wisconsin Card Sorting Test; WAIS = Wechsler adult intelligence scale.
*Significant on a .0038 level (corrected for multiple comparisons)
A logistic regression (LR forward) was performed including all significant tests identified by ROC-analyses as predictors. The model significantly increased the number of correctly classified cases (χ2=17.841, p<0.001 with df=1). By chance, 79.5% of the cases would be classified correctly, and including the neuropsychological tests in the prediction improved the classification rate to 89.7%. The one variable that contributed significantly to the prediction was categorical fluency (Wald=6.821, Exp(B)=0.618, p=0.009). These results indicate that when a categorical fluency test score increases by one unit, the chance of an individual converting to the disease stage gets .618 times less likely. Therefore, a decrease in test scores increases the chances of converting to the disease stage. The R2 of .576
24
fluency test scores for converters and non-converters. The line at 19 shows the optimal cut-off score which yields highest sensitivity and specificity numbers.
Figure 4 No n-c on ve rte rs Co nv ert ers 0 1 0 2 0 3 0 4 0 C a te g o r ic a l fl u e n c y (R a w t e s t s c o r e s ) Discussion
The current study investigated presymptomatic cognitive decline in carriers of autosomal dominant gene mutations causing FTD. Eight mutation carriers converted to symptomatic FTD within the study time window. Cognitive performances declined over time in converters within the domains of social cognition, attention and mental processing speed, executive function, language, and memory. Within the group of converters, we found gene-specific patterns of decline in MAPT and GRN mutation carriers, with MAPT converters showing lower scores on tests of social cognition, attention and mental processing speed, executive function, memory, and language, and GRN converters declining in attention and mental processing speed, and executive function. Explorative analyses showed that the language measure categorical fluency was able to differentiate between converters and
non-25
converters at time of conversion. None of the tests were able to differentiate between converters and non-converters as early as four and two years before symptom onset. The current study offers starting points for the use of neuropsychological tests as cognitive biomarkers in FTD and confirms the value of neuropsychological assessment in facilitating early FTD diagnostics.
We found cognitive deterioration over time in MAPT mutation carriers within the domains of social cognition, language, and memory, but these findings did not remain significant once the converters were excluded from analyses. This observation refutes the idea of early presymptomatic cognitive decline and might indicate that cognitive functions are not yet affected in mutation carriers during the early years before disease onset. These results show that the predictive value of neuropsychological assessment is limited in the presymptomatic phase. Nevertheless, these effects could be influenced by the limited sample sizes used in the current study. Secondly, there is a possibility that the majority of mutation carriers in the current study were far from symptom onset, which could have weakened the effects. Our explorative analyses showed that neuropsychological tests are not able to correctly classify converters and non-converters up until the moment of conversion. Based on these results, there are signs that cognitive deterioration in mutation carriers is more of an explosive process, which starts somewhere within the last two years before conversion. This information is highly valuable, as uncovering neuropsychological tests that are able to reveal which patients are converting to the disease stage could facilitate diagnostic processes.
Lower performances on measures of social cognition, attention and mental
processing speed, executive function, and language in FTD-converters are in accordance with previous research, as is the sparing of visuoconstruction abilities (Neary, Snowden, & Mann, 2005; Rascovsky et al., 2011; Seelaar, Rohrer, Pijnenburg, Fox, & van Swieten, 2010). Specifying our group of converters into MAPT and GRN converters for explorative purposes
26
showed partly overlapping but mostly unique profiles of cognitive decline. MAPT converters showed decline over time within the domains of social cognition, attention and mental processing speed, executive function, and language, which confirms previous research (Jiskoot et al., 2016; Rascovsky et al., 2011; Seelaar, Rohrer, Pijnenburg, Fox, & van Swieten). In addition, our study found significant decline over time on memory measures in MAPT converters. As for GRN converters, we found the domains of attention and mental processing speed, and executive function to decline over time, which was also shown by Barandiaran et al. (2012) in the early disease phase. Contrary to Barandiaran et al. (2012), we did not find significant decline for language/naming tests in GRN converters. Possibly we did not find as many significant results for GRN converters as for MAPT converters due to the smaller sample of GRN converters.
In an attempt to identify neuropsychological tests that are able to differentiate between converters and non-converters, we found that tests within the domains of social cognition, executive function and language were able to differentiate between converters and non-converters at the moment of conversion. Categorical fluency showed the most promising results, being able to correctly classify converters and non-converters and having both satisfactory sensitivity and specificity indexes. The other tests showed some ability to distinguish between converters and non-converters, but meanwhile included a considerable rate of false positives. These tests are thus not ideal predictors and are not to be used in clinical practice individually. As for categorical fluency, converters in a group of mutation carriers were correctly identified in 87.5% of the cases, and non-converters were diagnosed correctly in 93.3% of the cases. In terms of negative and positive predictive value, this means that categorical fluency is able to state with 96.9% certainty that a patient who scores above the cut-off of 19 is not developing FTD, and once a patient scores lower than the cut-off, it can be stated that this patient is developing FTD with 77.8% certainty. In medical settings high sensitivity is generally preferred over high specificity (Hendriks et al., 2014). In clinical
27
practice this translates to not taking the risk to give a patient a negative diagnosis when they actually are ill, and therefore accepting a higher rate of false positives. In our case however, categorical fluency is more specific than sensitive, which means that it more often gives a correct negative diagnosis, than a correct positive diagnosis. Categorical fluency could therefore be an efficient measure to include in screenings for patients who are suspected of developing FTD, as it can very effectively rule out an FTD syndrome in a substantial
proportion of the cases. As a useful first screening instrument, categorical fluency has the advantages of being easily applicable in clinical practice; it is of low costs, it does not require complex test material, instructions are straightforward and it is currently already included in a great majority of test batteries. However, since a substantial number of cases will be given a false positive diagnosis, it should be insisted that the diagnostic process remains
multidisciplinary by merging clinical impressions, neuropsychological performances and MRI data a.o. in the diagnostic process. Furthermore, other neuropsychological tests should be included in order to obtain a reliable impression of cognitive performances.
Neuropsychological deterioration in symptomatic FTD has been extensively described in the literature, but current findings can be of additional use in early FTD diagnosis. We are familiar with cognitive problems in FTD, but they are not always recognized in the early disease phase, causing a diagnostic delay (Hodges et al., 2003). The use of cognitive biomarkers is therefore desirable in order to facilitate early diagnostic processes.
The literature shows that fluency measures have previously been pointed out for their use in diagnosing Alzheimer’s disease and other neurodegenerative disorders (Zhao et al., 2013). In a paper by Pakhomov and Hemmy (2014), it was shown that lower Fluency scores correlate with memory impairment and a significantly higher risk for developing Alzheimer’s disease (AD). Furthermore, semantic fluency measures have been shown to be impaired in patients with focal lesions in temporal cortex areas (Henry & Crawford, 2004; Henry, Crawford & Phillips, 2004). These results and results from the current study suggest a
28
possible role for categorical fluency as a cognitive biomarker for FTD. Previous studies addressed neuroimaging measures and levels of neurofilament light chain (NFl) in
cerebrospinal fluid as possible biomarkers in FTD (Dopper et al., 2012; Meeter et al., 2016; Rohrer et al., 2015), and cognitive biomarkers could be a valuable addition to the diagnostic process. Categorical fluency has now been identified as a potential candidate, since the current results suggest that categorical fluency is able to distinguish converters from non-converters at symptom onset. More research and validation studies could confirm the value of this test as a cognitive biomarker, and it would be interesting to assess the value of categorical fluency in tracking disease progression.
The finding that categorical fluency is the only test that reaches statistical
significance throughout all analyses can possibly be explained by the larger number of MAPT converters than GRN converters in the sample, causing MAPT converters to largely drive the effects. Mutations in MAPT genes are strongly associated with language dysfunctions— specifically semantic deterioration (Seelaar, Rohrer, Pijnenburg, Fox, & van Swieten, 2010)— that is the construct measured by categorical fluency. Perhaps replicating the analyses within a larger group of GRN converters would identify other tests with differentiating capacities.
Where MAPT and GRN mutation carriers showed patterns of cognitive decline, healthy controls showed improvement on neuropsychological tests over time on a variety of tests within the domains of social cognition and memory. Learning effects are not
unexpected, since repeated measures research designs are basically a form of training resulting in improved test performances (Beglinger et al., 2005). On the contrary, no such learning effects were observable in MAPT mutation carriers and GRN mutation carriers, who deteriorated on the same tests. Where healthy controls show effects of learning on some of the tests, carriers of a mutated MAPT or GRN gene lack these learning curves. The absence of learning effects is highly informative, since not only deterioration but also stagnation
29
patterns in test performances are relevant when improvement patterns are expected. The finding of memory decline in converters strengthens the hypotheses about memory functions getting compromised in FTD. Based on the results of the current study and on recent findings in literature, memory dysfunctions should not necessarily be considered exclusion criteria for FTD diagnoses (Hornberger & Piguet, 2012; Jiskoot et al., 2016). The causes of memory decline remain speculative, but previous research suggests that the cause is not a pure memory dysfunction, and can be explained by deterioration of executive functions like active learning strategies and information retrieval processes (Collette et al., 2010). According to Hornberger and Piguet (2012), more sensitive memory measures should be included in future neuropsychological test batteries, which should be capable of differentiating between processes relying on prefrontal cortex structures and medial temporal lobe structures, which are essential for executive function and memory, respectively.
Even though the current study elaborated on the study by Jiskoot et al. (2016), there are differences between outcomes of the current study on the one hand and the paper by Jiskoot et al. (2016) on the other hand. The latter lacked information about actual ages at disease onset and used a model with estimates of ages of onset instead. In the current study the number of converters substantially increased compared to the study by Jiskoot et al. (2016), which gave us factual information about ages of onset. As a consequence, the current study offered a model by which a more accurate picture of the presymptomatic phase could be drawn.
Strong aspects of the study were the longitudinal design in which we followed large families with MAPT or GRN gene mutations for multiple years. Where other studies used cross-sectional designs, the current study performed true longitudinal analyses, which made both between group comparisons as within subject comparisons possible. Elaborating on this point, the additional follow-up assessment included in the current study offered a more
30
robust research design which was able to acquire more information altogether, creating a clearer picture of the disease processes in FTD. Thirdly, as mentioned before, the current study no longer needed to use estimated ages of onset in the analyses, but acquired information about actual ages of onset. Lastly, we included a wide range of
neuropsychological tests in our study, in order to cover a large set of cognitive functions. FTD research did not yet identify a congruent profile of cognitive functions compromised in FTD, and assessing an extensive amount of cognitive functions could contribute to reaching a consensus on what to expect from FTD patients in clinical practice.
A limitation of the current study was the small number of converters within our sample. Due to these small sample sizes, current findings should be interpreted with caution. Future research with larger numbers of converters should confirm or extend current findings. In these samples, it would be informative to split the sample further into cases of bvFTD and PNFA. In the current sample, these separate cases would nearly create case studies of which the power criterion could not be justified. Secondly, since conversion status is partly based on neuropsychological test performances, the fact that we investigated neuropsychological decline in converters is prone to circular argumentation. Converters are likely to show declined performances on neuropsychological tests, since these patterns of decline were part of their diagnostic processes. However, it should be put in perspective that conversion status is based on a variety of parameters, including clinical impressions, patient history, caregiver anamnesis, cognitive screeners, MRI, and lumbar punctures.
In conclusion, we see subtle presymptomatic decline on group level in the last two years before disease onset in mutation carriers of autosomal dominant FTD genes. We see widely scattered profiles of cognitive decline in converters, from language functions extending to social cognition, attention and mental processing speed, executive function, and memory, with specific decline patterns in MAPT and GRN gene mutations.
31
fluency. The early disease stage seems unsuitable for predictions of the development of FTD, but cognitive biomarkers could be of use within the last two years before disease onset and after in tracking the onset and progression of FTD disease processes. The current study offered starting points for the identification of cognitive biomarkers, which would be highly valuable additions to current diagnostic processes, by facilitating early diagnostics and therefore eliminating diagnostic delay in FTD.
REFERENCES
Barandiaran, M., Estanga, A., Moreno, F., Indakoetxea, B., Alzualde, A., Balluerka, N., & de Munain, A. L. (2012). Neuropsychological features of asymptomatic c. 709-1G> A progranulin mutation carriers. Journal of the International Neuropsychological Society, 18(06), 1086-1090.
Beck, A. T., Steer, R. A., & Brown, G. K. (1996). Beck depression inventory.
Beglinger, L. J., Gaydos, B., Tangphao-Daniels, O., Duff, K., Kareken, D. A., Crawford, J., ... & Siemers, E. R. (2005). Practice effects and the use of alternate forms in serial neuropsychological testing. Archives of Clinical Neuropsychology, 20(4), 517-529. Bondi, M. W., Serody, A. B., Chan, A. S., Eberson-Shumate, S. C., Delis, D. C., Hansen, L. A., &
Salmon, D. P. (2002). Cognitive and neuropathologic correlates of Stroop Color-Word Test performance in Alzheimer's disease. Neuropsychology, 16(3), 335.
Bouma, A., Mulder, J., Lindeboom, J., & Schmand, B. (Eds.). (2012). Handboek neuropsychologische diagnostiek. Amsterdam: Pearson.
Chaytor, N., Schmitter-Edgecombe, M., & Burr, R. (2006). Improving the ecological validity of executive function assessment. Archives of clinical neuropsychology, 21(3), 217-227.
Chow, T. W., Miller, B. L., Hayashi, V. N., & Geschwind, D. H. (1999). Inheritance of frontotemporal dementia. Archives of Neurology, 56(7), 817-822.\
32
Collette, F., Van der Linden, M., & Salmon, E. (2010). Dissociation between controlled and automatic processes in the behavioral variant of fronto-temporal dementia. Journal of Alzheimer's Disease, 22(3), 897-907.
Cummings, J. L., Mega, M., Gray, K., Rosenberg-Thompson, S., Carusi, D. A., & Gornbein, J. (1994). The Neuropsychiatric Inventory comprehensive assessment of
psychopathology in dementia. Neurology, 44(12), 2308-2308.
Cummings, J. L. (1997). The Neuropsychiatric Inventory Assessing psychopathology in dementia patients. Neurology, 48(5 Suppl 6), 10S-16S.
Dopper, E. G., Chalos, V., Ghariq, E., den Heijer, T., Hafkemeijer, A., Jiskoot, L. C., & Rombouts, S. A. (2016). Cerebral blood flow in presymptomatic MAPT and GRN mutation carriers: A longitudinal arterial spin labeling study. NeuroImage: Clinical, 12,
460-465.
Dopper, E. G., Rombouts, S. A., Jiskoot, L. C., den Heijer, T., de Graaf, J. R. A., de Koning, I., & van Buchem, M. A. (2013). Structural and functional brain connectivity in
presymptomatic familial frontotemporal dementia. Neurology, 80(9), 814-823. Douglas, V. I. (1972). Stop, look and listen: The problem of sustained attention and
impulse control in hyperactive and normal children. Canadian Journal of
Behavioural Science/Revue canadienne des sciences du comportement, 4(4), 259. Dubois, B., Slachevsky, A., Litvan, I., & Pillon, B. F. A. B. (2000). The FAB A frontal assessment
battery at bedside. Neurology, 55(11), 1621-1626.
Ekman, P., & Friesen, W. (1976). Pictures of facial affect. Palo Alto, CA: Consulting Psychologists Press.
Folstein, M. F., Folstein, S. E., & McHugh, P. R. (1975). “Mini-mental state”: a practical method for grading the cognitive state of patients for the clinician. Journal of psychiatric research, 12(3), 189-198.
33
Grant, D. A., & Berg, E. (1948). A behavioral analysis of degree of reinforcement and ease of shifting to new responses in a Weigl-type card-sorting problem. Journal of
experimental psychology, 38(4), 404.
Hachioui, H. E., van de Sandt-Koenderman, M. W., Dippel, D. W., Koudstaal, P. J., & Visch-Brink, E. G. (2012). The ScreeLing: occurrence of linguistic deficits in acute aphasia post-stroke. Journal of rehabilitation medicine, 44(5), 429-435.
Happé, F., Brownell, H., & Winner, E. (1999). Acquired ‘theory of mind' impairments following stroke. Cognition, 70(3), 211-240.
Harrison, J. E., Buxton, P., Husain, M., & Wise, R. (2000). Short test of semantic and phonological fluency: Normal performance, validity and test‐retest reliability. British Journal of Clinical Psychology, 39(2), 181-191.
Hendriks, M. P., Kessels, R. P. C., Gorissen, M. E. E., Schmand, B. A., & Duits, A. A. (2014). Neuropsychologische diagnostiek. De klinische praktijk. Amsterdam: Boom.
Henry, J. D., & Crawford, J. R. (2004). A meta-analytic review of verbal fluency performance following focal cortical lesions.
Henry, J. D., Crawford, J. R., & Phillips, L. H. (2004). Verbal fluency performance in dementia of the Alzheimer’s type: a meta-analysis. Neuropsychologia, 42(9), 1212-1222. Hodges, J. R., Davies, R., Xuereb, J., Kril, J., & Halliday, G. (2003). Survival in frontotemporal
dementia. Neurology, 61(3), 349-354.
Hornberger M., Piguet O. Episodic memory in frontotemporal dementia: a critical review. Brain 2012;135:678–692.
Irwin D.J., Brettschneider J., McMillan C.T., et al. Deepclinical and neuropathological phenotyping of Pick disease.Ann Neurol 2016;79:272–287.
Jensen, A. R. (1965). Scoring the Stroop test. Acta Psychologica, 24, 398-408. Jensen, A. R., & Rohwer, W. D. (1966). The Stroop color-word test: a review. Acta
34
Jiskoot, L. C., Dopper, E. G., den Heijer, T., Timman, R., van Minkelen, R., van Swieten, J. C., & Papma, J. M. (2016). Presymptomatic cognitive decline in familial frontotemporal dementia A longitudinal study. Neurology, 87(4), 384-391.
Kaplan, E. (1983). The assessment of aphasia and related disorders (Vol. 2). Lippincott Williams & Wilkins.
Kaufer, D. I., Cummings, J. L., Ketchel, P., Smith, V., MacMillan, A., Shelley, T., & DeKosky, S. T. (2000). Validation of the NPI-Q, a brief clinical form of the Neuropsychiatric Inventory. The Journal of neuropsychiatry and clinical neurosciences, 12(2), 233-239. Lindeboom, J., Schmand, B., Tulner, L., Walstra, G., & Jonker, C. (2002). Visual association
test to detect early dementia of the Alzheimer type. Journal of Neurology, Neurosurgery & Psychiatry, 73(2), 126-133.
Marien, P., Mampaey, E., Vervaet, A., Saerens, J., & De Deyn, P. P. (1998). Normative data for the Boston Naming Test in native Dutch-speaking Belgian elderly. Brain and language, 65(3), 447-467.
Mathuranath, P. S., Nestor, P. J., Berrios, G. E., Rakowicz, W., & Hodges, J. R. (2000). A brief cognitive test battery to differentiate Alzheimer's disease and frontotemporal dementia. Neurology, 55(11), 1613-1620.
Meeter, L. H., Dopper, E. G., Jiskoot, L. C., Sanchez‐ Valle, R., Graff, C., Benussi, L., & Laforce Jr, R. (2016). Neurofilament light chain: a biomarker for genetic frontotemporal dementia. Annals of Clinical and Translational Neurology, 3(8), 623-636.
Meeter, L. H., Patzke, H., Loewen, G., Dopper, E. G., Pijnenburg, Y. A., van Minkelen, R., & van Swieten, J. C. (2016). Progranulin Levels in Plasma and Cerebrospinal Fluid in Granulin Mutation Carriers. Dementia and Geriatric Cognitive Disorders Extra, 6(2), 330-340.
Mitchell, A. J. (2013). The Mini-Mental State Examination (MMSE): an update on its
diagnostic validity for cognitive disorders. In Cognitive Screening Instruments (pp. 15-46). Springer London.
35
Nagelkerke, N. J. (1991). A note on a general definition of the coefficient of determination. Biometrika, 78(3), 691-692.
Neary, D., Snowden, J., & Mann, D. (2005). Frontotemporal dementia. The Lancet Neurology, 4(11), 771-780.
Pakhomov, S. V., & Hemmy, L. S. (2014). A computational linguistic measure of clustering behavior on semantic verbal fluency task predicts risk of future dementia in the Nun Study. Cortex, 55, 97-106.
Paul, F., Erdfelder, E., Buchner, A., & Lang, A. G. (2009). Statistical power analyses using G* Power 3.1: Tests for correlation and regression analyses. Behavior research methods, 41(4), 1149-1160.
Pickering-Brown, S. M., Rollinson, S., Du Plessis, D., Morrison, K. E., Varma, A., Richardson, A. M., & Mann, D. M. (2008). Frequency and clinical characteristics of progranulin mutation carriers in the Manchester frontotemporal lobar degeneration cohort: comparison with patients with MAPT and no known mutations. Brain, 131(3), 721-731.
Puente, A. (1985). Wisconsin card sorting test. Test critiques, 4, 677-682.
Rascovsky, K., Hodges, J. R., Knopman, D., Mendez, M. F., Kramer, J. H., Neuhaus, van Swieten, J.C., Seelaar, H., Dopper, G. P., Onyike, C. U. & Hillis, A. E. (2011). Sensitivity of revised diagnostic criteria for the behavioural variant of frontotemporal dementia.
Brain, 134(9), 2456-2477.
Ratnavalli, E., Brayne, C., Dawson, K., & Hodges, J. R. (2002). The prevalence of frontotemporal dementia. Neurology, 58(11), 1615-1621.
Reitan, R. M. (1992). TMT, Trail Making Test A & B. South Tucson, AR: Reitan Neuropsychology Laboratory.
Rohrer, J. D., Nicholas, J. M., Cash, D. M., van Swieten, J., Dopper, E., Jiskoot, L., & Espak, M. (2015). Presymptomatic cognitive and neuroanatomical changes in
36
genetic frontotemporal dementia in the Genetic Frontotemporal dementia Initiative (GENFI) study: a cross-sectional analysis. The Lancet Neurology, 14(3), 253-262.
Rohrer, J. D., Ridgway, G. R., Modat, M., Ourselin, S., Mead, S., Fox, N. C., & Warren, J. D. (2010). Distinct profiles of brain atrophy in frontotemporal lobar degeneration caused by progranulin and tau mutations. Neuroimage, 53(3), 1070-1076. Rosenberg, S. J., Ryan, J. J., & Prifitera, A. (1984). Rey auditory verbal learning test performance of patients with and without memory impairment. Journal of clinical psychology, 40(3), 785-787.
Royall, D. R., Cordes, J. A., & Polk, M. (1998). CLOX: an executive clock drawing task. Journal of Neurology, Neurosurgery & Psychiatry, 64(5), 588-594.
Schmidt, M. (1996). Rey auditory verbal learning test: A handbook (p. 1996). Los Angeles: Western Psychological Services.
Seelaar, H., Rohrer, J. D., Pijnenburg, Y. A., Fox, N. C., & Van Swieten, J. C. (2010). Clinical, genetic and pathological heterogeneity of frontotemporal dementia: a review. Journal of Neurology, Neurosurgery & Psychiatry, jnnp-2010.
Smith, A. (1973). Digit symbol modalities test. Los Angeles: Western Psychological Services. Snowden, J. S., Adams, J., Harris, J., Thompson, J. C., Rollinson, S., Richardson, A., &
Pickering-Brown, S. (2015). Distinct clinical and pathological phenotypes in frontotemporal dementia associated with MAPT, PGRN and C9orf72 mutations. Amyotrophic Lateral Sclerosis and Frontotemporal Degeneration, 16(7-8), 497-505. Storch, E. A., Roberti, J. W., & Roth, D. A. (2004). Factor structure, concurrent validity, and
internal consistency of the beck depression inventory—second edition in a sample of college students. Depression and anxiety, 19(3), 187-189.
Squire, L.R. (1982). The neuropsychology of human memory. Annual Review of Neuroscience, 5, 52-64.
37
Stroop, J. (1935). Studies of interference in serial verbal reactions. Journal of Experimental Psychology, 18, 643-662.
Tombaugh, T. N., & McIntyre, N. J. (1992). The mini‐ mental state examination: a
comprehensive review. Journal of the American Geriatrics Society, 40(9), 922-935. Troyer, A. K., Moscovitch, M., & Winocur, G. (1997). Clustering and switching as two
components of verbal fluency: evidence from younger and older healthy adults. neuropsychology, 11(1), 138.
Troyer, A. K., Moscovitch, M., Winocur, G., Alexander, M. P., & Stuss, D. (1998). Clustering and switching on verbal fluency: The effects of focal frontal-and temporal-lobe lesions. Neuropsychologia, 36(6), 499-504.
Van der Elst, W., van Boxtel, M. P., van Breukelen, G. J., & Jolles, J. (2006). The Letter Digit Substitution Test: normative data for 1,858 healthy participants aged 24–81 from the Maastricht Aging Study (MAAS): influence of age, education, and sex. Journal of clinical and experimental neuropsychology, 28(6), 998-1009.
Visch-Brink, E. G., Sandt-Koenderman, W. M. E., & el Hachioui, H. (2010). ScreeLing: handleiding. Bohn Stafleu van Loghum.
Visch-Brink, E. G., Stronks, D. L., & Denes, G. (2005). De semantische associatie test. Harcourt Test Publishers.
Warren, J. D., Rohrer, J. D., & Rossor, M. N. (2013). Frontotemporal dementia. Bmj, 347, f4827.
Wechsler, D. (1997). WAIS-III: Wechsler adult intelligence scale. San Antonio, TX: Psychological Corporation.
Zhao, Q., Guo, Q., & Hong, Z. (2013). Clustering and switching during a semantic verbal fluency test contribute to differential diagnosis of cognitive impairment. Neuroscience bulletin, 29(1), 75-82.
38
Appendix AA. Sample characteristics: global cognitive functioning and psychiatric screeners
Global cognitive functioning
The Mini Mental State Examination (MMSE) (Folstein, Folstein, & McHugh, 1975) and the Frontal Assessment Battery (Dubois, Slachevsky, Litvan, & Pillon, 2000) were used as
screeners for cognitive functioning and frontal lobe functioning, respectively. The MMSE is a short test battery comprising 20 items covering 11 different domains. A higher score
indicates better performance, with a maximum of 30 points. Any score under the 24-point cut-off may indicate cognitive impairment (Tombaugh & McIntyre, 1992). According to Mitchell (2013), the MMSE is a well-validated and reliable measure. The FAB consists of six subtests with a maximum score of 3, making the maximum total score 18. The FAB has good inter-rater reliability and internal consistency with a .87 Kappa and .78 Cronbach’s Alpha. Discriminant validity is high (89.1%) (Dubois, Slachevsky, Litvan, & Pillon, 2000).
Psychiatric screeners
The Neuropsychiatric Inventory. The neuropsychiatric inventory (NPI) (Cummings et al., 1994) is a questionnaire which assesses different neuropsychiatric disturbances in dementia
patients. The questionnaire is filled out by a close caregiver of the patient. The questionnaire consists of 20 items covering 20 neuropsychiatric symptoms. The caregiver is instructed to indicate the presence of certain symptoms, its frequency, severity and emotional burden on the caregiver (Cummings, 1997). The outcome measure is the raw total score, which gives an indication of the presence of neuropsychiatric symptoms in the participant. Test-retest reliability and convergent validity are satisfactory (Kaufer et al., 2000).
Beck Depression Inventory (BDI). The BDI is a widely used screener for depressive symptoms (Beck, Steer, & Brown, 1996). The participant is instructed to report the presence of
depressive symptoms in the past week. The 21 items can be answered by four possibilities, with a maximum score of 63. Where answer A infers absence of the symptom and is worth
