Applied Neuropsychology: Adult

ISSN: 2327-9095 (Print) 2327-9109 (Online) Journal homepage: [Link]

Effects of Korsakoff Amnesia on performance and

symptom validity testing

Erik Oudman, Emmy Krooshof, Roos van Oort, Beth Lloyd, Jan W. Wijnia &
Albert Postma

To cite this article: Erik Oudman, Emmy Krooshof, Roos van Oort, Beth Lloyd, Jan W. Wijnia &
Albert Postma (2019): Effects of Korsakoff Amnesia on performance and symptom validity testing,
Applied Neuropsychology: Adult, DOI: 10.1080/23279095.2019.1576180

To link to this article: [Link]

Published online: 08 Mar 2019.

Submit your article to this journal

Article views: 24

View Crossmark data

Full Terms & Conditions of access and use can be found at

[Link]
APPLIED NEUROPSYCHOLOGY: ADULT
[Link]

Effects of Korsakoff Amnesia on performance and symptom validity testing

Erik Oudmana,b , Emmy Krooshofa,b, Roos van Oortc, Beth Lloyda,b, Jan W. Wijniaa,b, and
Albert Postmaa,b
a
Department of Experimental Psychology, Helmholtz Institute, Utrecht University, Utrecht, The Netherlands; bSlingedael Korsakoff
Center, Lelie Care Group, Rotterdam, The Netherlands; cWettstein & Peterse Expertise (WPEX), Amersfoort, The Netherlands

ABSTRACT KEYWORDS
Performance validity tests (PVTs) and Symptom validity tests (SVTs) are developed to identify Korsakoff; neuropsycho-
people that present false or exaggerated symptoms. Although a key factor of both types of logical assessment;
tests includes relative insensitivity to cognitive disorders, the direct effects of amnesia have performance validity test;
Test of Memory
been poorly studied. Therefore, a sample of 20 patients diagnosed with Korsakoff Amnesia Malingering;
(KA) through neuropsychological assessment and 20 healthy comparisons (HC) were admin- underperformance
istered the Test of Memory Malingering (TOMM), the Structured Inventory of Malingered
Symptomatology (SIMS), and the newly developed Visual Association Test – Extended (VAT-
E). Our results show that KA patients scored systematically lower on the TOMM and VAT-E,
while performance on the SIMS was comparable with healthy comparisons. Some KA
patients were regarded as underperformers based on the TOMM and VAT-E, suggesting limi-
tations in applying these instruments in severe amnesia. There was a strong interdepend-
ence of PVTs in logistic regression. We conclude that the TOMM and VAT-E are not fully
robust against severe memory disorders and show a serious risk of false positives. Complete
neuropsychological profile analysis is needed, and PVTs should be interpreted with caution
in patients with suspected amnesia.

Introduction PVTs are standalone or embedded tests designed and

administered for the purpose of determining invalid
When using cognitive tests for diagnostic purposes it
responding. At face value, they seem to be similar to
is crucial to know whether the patient is exaggerating
or fabricating cognitive deficits. Between 22 and 40% other neuropsychological tests, although they are
of all the patients involved in litigation have noncredi- largely insensitive to brain impairments. Practically
ble performance on neuropsychological assessments speaking, cognitively impaired patients are able to
(Mittenberg, Patton, Canyock, & Condit, 2002; pass these tests despite the severity of their cognitive
Rogers, 2008). Performance validity tests (PVTs) are problems (Wodushek & Greher, 2017). Some PVTs
tests for assessing the credibility of the patient’s per- have fixed cutoff scores, but others are based on hier-
formance on neuropsychological assessments, and archical profile analysis of the PVT in relation to
symptom validity tests (SVTs) are questionnaires to other neuropsychological tests. For example, in the
index exaggeration of symptoms (Bush et al., 2005). Medical Symptom Validity Test, performance on
Failure on a PVT or SVT reflects contributions to test “easy” and “hard” subtests is directly compared, to
performance other than neurological or cognitive fac- reduce false positives (Wodushek & Greher, 2017).
tors (Bigler, 2012). This is particularly the case when SVTs are self-report questionnaires that are designed
memory symptoms are feigned or effort was too little. to detect feigned or exaggerated psychopathology. The
Recently neuropsychologists have become increas- general idea is that respondents are likely to endorse
ingly aware that, although cognitive testing is highly bizarre, rare, atypical, or extreme symptoms on a
sensitive to brain dysfunction, poor test results are questionnaire when they attempt to feign or exagger-
nonspecific and can arise for a variety of reasons ate symptoms (Van Impelen, Merckelbach, Jelicic, &
(Greher & Wodushek, 2017). Not surprisingly, the use Merten, 2014).
of PVTs and SVTs has climbed dramatically in clinical Many PVTs and SVTs have been developed and
practice (Sweet, Benson, Nelson, & Moberg, 2015). validated over the years. The most frequently applied,

CONTACT Erik Oudman [Link]@[Link] Experimental Psychology, Helmholtz Institute, Utrecht University, Heidelberglaan 1, 3584CS
Utrecht, The Netherlands.
ß 2019 Taylor & Francis Group, LLC
2 E. OUDMAN ET AL.

and probably the most validated PVT is the TOMM cognitive disorders on PVT performance is that cogni-
(Dandachi-FitzGerald, Ponds, & Merten, 2013; tive functioning is directly related to PVT perform-
Martin, Schroeder, & Odland, 2015). The TOMM is ance in some studies on dementia. Rudman et al.
based on visual stimuli where the patient has to (2011) found that specifically learning of new infor-
remember and recognize 45 out of 50 individual pic- mation was significantly correlated with task perform-
tures (Tombaugh, 1996). Since the TOMM has ance on the frequently applied Test of Memory
become a very popular PVT, with even online tutori- Malingering (TOMM), and the Medical Symptom
als how to pass this test as a malingerer (Wodushek & Validity Test. Selective attention was also significantly
Greher, 2017), there is an urgent need to develop new related to the Coin in hand test. The authors argued
PVTs. Moreover, in new PVTs it would be relevant to that the specificity of the applied PVTs was not opti-
combine assessment of declarative memory, and valid- mal, since some moderate to severe dementia patients
ity testing in one test. Recently, the Visual Association were classified as displaying suboptimal effort. Bortnik
Test-Extended (VAT-E) was developed as a new PVT et al. (2013) made even stronger claims, stating that
that is also able to detect declarative memory prob- the majority of effort tests demonstrated unacceptably
lems in subtests. The VAT-E forms an extension of high false-positive rates in their sample of 164 demen-
the traditionally used Visual Association Test (VAT) tia patients. In fact, 22% of the patients were unable
which is an episodic memory test (Lindeboom, to pass the second trial of the TOMM, being misclas-
Schmand, & Christensen, 2003; Meyer, de Jonghe, sified as underperformers. In contrast, some norma-
Schmand, & Ponds, 2017), and yields very promising tive data suggest that the effects of cognitive disorders
results in discriminating patients with Alzheimer’s dis- are rather limited (Tombaugh, 1997).
ease, Mild Cognitive Impairment (MCI), and healthy Although PVTs are widely used in a variety of set-
persons that feign memory deficits (Meyer et al., tings to differentiate feigned and actual memory
2017). In the VAT-E both fixed cutoff scores are com- issues, there are currently very limited studies avail-
bined with hierarchical profile analysis, comparing able on the validity of PVTs in patients with profound
scores on multiple indices. A specifically prominent memory disorders with relatively preserved other
SVT that is developed is the Structured Inventory of functions. Earlier investigations on PVT performance
Malingered Symptomatology (SIMS) (Smith & Burger, in amnesia frequently incorporated patients in which
1997). In contrast to PVTs, the SIMS is not developed memory was not the only or central cognitive prob-
to indicate underperformance on neuropsychological lem, but part of general cognitive dysfunction (Loring,
assessments, but over-reporting of symptomatology. Larrabee, Lee, & Meador, 2007; Tombaugh, 1997). In
Dandachi-FitzGerald, Ponds, Peters, and Merckelbach the present study, we enrolled patients with Korsakoff
(2011) emphasized on the importance to test both Amnesia (KA). KA is a chronic brain disorder, caused
aspects of symptom exaggeration, since over-represen- by thiamin (vitamin B1) deficiency. In the industrial-
tation of symptoms was not necessarily related to ized world, the most common cause of thiamin defi-
underperformance. A thorough assessment of validity ciency is severe alcoholism. The most essential
therefore needs to take both dimensions into account. symptom of KA is profound declarative amnesia for
An important question is whether PVTs and SVTs learning and remembering new material (Kopelman,
are able to discriminate patients with actual cognitive Thomson, Guerrini, & Marshall, 2009). Commonly,
deficits from patients that feign symptoms. Currently, but not necessarily, executive deficits are present, such
the literature is inconclusive on this matter, with as problems with inhibition of behavior, high interfer-
some studies suggesting that the majority of PVTs are ence of information sensitivity, poor judgment, poor
not always suitable for patients with severe cognitive planning abilities, problem solving inabilities, and per-
deficits (Bortnik, Horner, & Bachman, 2013; severative responses (Moerman - van den Brink et al.,
Dorociak, Schulze, Piper, Molokie, & Janecek, 2018; 2019). In contrast to many forms of dementia,
Rudman, Oyebode, Jones, & Bentham, 2011; Walter, apraxia, aphasia, agnosia, and broad intellectual
Morris, Swier-Vosnos, & Pliskin, 2014, see McGuire, decline are not symptoms of KA (Kopelman et al.,
Crawford, & Evans, 2019 for a review), whereas other 2009), increasing the relevance of testing the sensitiv-
studies find that PVTs can be safely used in a popula- ity of PVTs and SVTs in KA.
tion with severe cognitive deficits (Hampson, Kemp, The aim of this study was to test whether patients
Coughlan, Moulin, & Bhakta, 2014; Liu et al., 2016; with severe amnesia are able to successfully pass the
Meyer et al., 2017; Slick et al., 2003; Tombaugh, 1996, TOMM, SIMS, and VAT-E. Moreover, we wanted to
1997). A point of criticism regarding the influence of elucidate how the outcome of the TOMM, SIMS, and
 APPLIED NEUROPSYCHOLOGY: ADULT 3

Table 1. Demographic characteristics.

Healthy Comparisons (n ¼ 20) Korsakoff Amnesia (n ¼ 20)
Number of females 12 10 v2 (1, n ¼ 40)¼ 0.40, p ¼ .53
Agea (M, SD) 55.4 (4.9) 57.8 (7.3) t(38) ¼ 1.14, p ¼ .260, g2 ¼ .04
Premorbid IQ scores (M, SD)b 97.7 (10.9) 93.6 (11.4) t(38) ¼ 1.22, p ¼ .230, g2 ¼ .04
VAT-E Paired Association (M, SD)c 16.6 (5.6) 5.6 (4.3) t(38) ¼ 6.77, p < .001, g2 ¼ .55
VAT-E Free Recall (M, SD)c 23.8 (6.6) 9.0 (5.2) t(38) ¼ 1.14, p < .001, g2 ¼ .62
CAMDEX Nonmemory impaired %d 21.4
CAMDEX Memory impaired %d 100.0
MMSE impaired %e 73.3
Digit Test Forward impaired %f 9.0
Digit Test Backward impaired %f 9.0
BADS impaired %g 25.0
a
Age in years.
b
Premorbid IQ scores were estimated with the Dutch version of the National Adult Reading Test (Schmand et al.,1991).
c
VAT-E subtest Paired Association and Free Recall are the memory measures of the VAT-E (Meyer & de Jonghe, 2017). Higher scores represent better per-
formance on the subtest.
d
CAMDEX total scores for the nonmemory and memory section, as an index of general cognitive functioning, normed according to Roth et al. (1986).
e
MMSE total scores, as an index of general cognitive functioning, normed according to Folstein et al. (1975).
f
Digit span, as a test for concentration and working memory, normed according to Lindeboom and Matto (1994).
g
Behavioral Assessment of the Dysexecutive Syndrome (BADS), a test battery for executive functioning, normed according to Wilson et al. (1996).

VAT-E could predict the presence of KA, and whether also enrolled in the study. IQ was estimated with the
combining evidence would reduce the effects of Dutch version of the National Adult Reading Test
severe amnesia. (Schmand, Bakker, Saan, & Louman, 1991). The pro-
ject was conducted according to the declaration of
Helsinki and written informed consent was obtained
Materials and methods
for all participants.
Participants
Twenty patients diagnosed with Korsakoff Amnesia Tasks and procedure
(KA) through extensive multidisciplinary observation
Test of Memory Malingering (TOMM)
and diagnostics participated in this study (see Table 1
In the TOMM, 50 pictures were shown sequentially to
for an overview). They were all inpatients of the
the participant. After the learning trial, the participant
Korsakoff Center “Slingedael,” Lelie Care Group, in
was presented with 50 forced choices between two
Rotterdam, The Netherlands. All patients fulfilled the
pictures of which one of the two had been shown in
DSM-5 criteria for the Alcohol-induced major neuro- the learning trial. After the first trial, the participant
cognitive disorder, Amnestic Confabulatory type was shown the same pictures in a second learning
(code: 291.1) (American Psychiatric Association, trial. Lastly, the participant was asked once more to
2013), and the characteristics of KA described by select the picture he/she had seen during the learning
Kopelman (2002). All patients had been sober for trial in 50 consecutive forced choices. During all trials
more than six months and required intensive sheltered the participant received feedback on their answer, as
living due to the severity of their amnesia. All patients indicated by the TOMM manual. TOMM trial 2
fulfilled the D criterion of Slick et al. (2003) regarding was analyzed.
the severity of genuine cognitive pathology and war-
ranted intensive sheltered living based on the severity Visual Association Test-Extended (VAT-E)
of their amnesia. The amnestic syndrome was con- For the VAT-E, the participant was instructed to look
firmed by extensive neuropsychological testing. All at and remember 24 pictures with one object, animal,
patients were in the chronic, amnestic stage of the or person on it, one by one for three seconds each, in
syndrome, none of the patients were in the confu- two learning trials. In the next section the 24 pictures
sional Wernicke psychosis at the moment of testing. were shown again, this time with an additional item
Other lifetime exclusion criteria were illiteracy, pres- in the picture. The participant was instructed to name
ence of additional neurological disorders (traumatic both items and identify which of the pictures had to
brain injury, epilepsy, stroke, or brain tumor), acute be remembered during the learning trials (immediate
psychiatric conditions (psychosis, major depression, recall; IR) to ensure the association of the pictures.
etc.), and physical conditions interfering with the test- After a 15-minute interval in which no memory or
ing procedure. Twenty healthy comparisons (HC), visual tests were administered, the participant was
statistically equivalent in age, IQ, and gender were again shown the 24 pictures with the additional
4 E. OUDMAN ET AL.

picture and was again asked to identify which picture the Dysexecutive Syndrome (BADS), a test battery for
had to be remembered during the learning trial executive functioning, was scored. A test score was con-
(delayed recall: DR). This was followed by the paired sidered “impaired” if the standard score was more than
association (PA) in which the participant was shown 1.5 SD below the normative mean) (Wilson, Alderman,
the initial 24 pictures and had to name which picture Burguess, Emslie, & Evans, 1996).
was added. Next, the participant was asked to freely
recall which of the pictures he/she still remembered
(free recall; FR). These could be named in pairs or Statistics
separately (e.g., either a candle on a football or a can-
The TOMM, VAT-E, and SIMS were presented in a
dle and a football). The last trial consisted of 12 of
the 24 pictures shown individually again as a cue and counterbalanced order. Because the assumptions for
the participant having to choose between four pictures parametric testing were violated by skewness of the
which was the associated picture and was interpreted data, nonparametric U-tests were performed on the
as the profile analysis score (FR  7 in combination TOMM trial 1 and trial 2, VAT-E IR, DR, and CN,
with MC  9). The consistency subtest (CN) was and SIMS, to index possible differences in median val-
derived from the IR and DR scores, and represented ues between KA patients and controls. Bonferroni cor-
whether the participant had the same answers correct rected p-values were presented. Moreover, Logistic
or wrong on the IR compared to the DR test. The regression was carried out as a hierarchical approach
subtests that measured underperformance were IR, to investigate accumulative prediction of severe
DR, and CN, and the profile analysis score. amnesia. Receiver Operated Characteristic curves were
estimated and objectified as the Area Under the Curve
Structured Inventory of Malingered (AUC) to index the ability of the tests to discriminate
Symptomatology (SIMS) KA patients and healthy controls correctly. Also,
The SIMS consisted of 75 items regarding symptom- Chi-squares were calculated for the proportions of
atology. If the amount of affirmative answers exceeded controls and patients that were unable to pass the
the cutoff value, it is an indication of feigning psychi- aforementioned tests, as well as the profile analysis
atric symptoms (Smith & Burger, 1997). A score score of the VAT-E, to indicate the effects of combin-
above 16 is indicative of aggravating. ing evidence.

Neuropsychological assessment
Results
Premorbid IQ scores were estimated with the Dutch
Table 1 shows a summary of neuropsychological tests,
version of the National Adult Reading Test for all par-
demographic variables, and the performance on back-
ticipants (Schmand et al., 1991). The VAT-E subtest
ground variables for both patients diagnosed with KA
Paired Association and Free Recall are the memory
and the healthy comparisons. Both groups were statis-
measures of the VAT-E (Meyer & de Jonghe, 2017)
tically equivalent on age, gender, and IQ. As expected,
and were collected for all participants. Higher scores
KA patients scored significantly lower on the VAT-E
represented better performance on the subtest.
memory indices, representing the severity of the
For the KA patients, the percentage of patients that
amnesia in KA (see Table 1).
scored below the cutoff value was reported for six add-
itional indices of cognitive functioning: as an index for
global cognitive functioning, the CAMDEX total scores
Prevalence of underperformance and over-
(cutoff value < 79 points), the nonmemory CAMDEX
reporting in Korsakoff Amnesia and healthy
score (cutoff < 62 points), and memory CAMDEX
comparisons
score (cutoff < 27 points) were reported (Roth et al.,
1986; Schmand, Walstra, Lindeboom, Teunisse, & Table 2 shows the median and range scores for the
Jonker, 2000). Moreover, the MMSE total scores (cut of patients diagnosed with KA and the healthy compari-
<24) was applied as an index of general cognitive func- sons. Task performance was significantly worse in KA
tioning (Folstein, Folstein, & McHugh, 1975). As a test patients for all trials of the TOMM and VAT-E, com-
for concentration and working memory, the digit span pared to healthy comparisons. The scores were com-
was applied (see Lindeboom & Matto, 1994 for differ- parable on the SIMS. Task performance on the PVTs,
ential norms). Finally, the Behavioural Assessment of but not the SVT, were sensitive for KA.
 APPLIED NEUROPSYCHOLOGY: ADULT 5

Table 2. Group differences across Performance and Symptom Validity Tests in 20 patients diagnosed with Korsakoff Amnesia and
20 Healthy comparisons.
Healthy Comparisons Korsakoff Amnesia Bonferroni Corrected
(n ¼ 20) (n ¼ 20) Statistics p-values
TOMM - Trial 1 (Median, Range)a 48.5 (41–50) 44.5 (24–49) U ¼ 79.50 p < .001
TOMM - Trial 2 (Median, Range)a 50.0 (49–50) 49.0 (42–50) U ¼ 76.50 p < .001
SIMS (Median, Range)b 4.0 (2–18) 6.0 (1–15) U ¼ 155.50 p ¼ 1.000
VAT-E Immediate Recognition (Median, Range)c 24.0 (23–24) 24.0 (21–24) U ¼ 118.50 p ¼ .002
VAT-E Delayed Recognition (Median, Range)c 24.0 (24) 22.5 (18–24) U ¼ 70,00 p < .001
VAT-E Consistency (Median, Range)c 24.0 (23–24) 23.0 (18–24) U ¼ 75.50 p < .001
Note. VAT-E ¼ Visual Association Test – Extended (Meyer et al., 2017); TOMM ¼ Test of Memory Malingering (Tombaugh, 1996); SIMS ¼ Structured
Inventory of Malingered Symptomatology (Smith & Burger, 1997).
a
Median Test score on the first and second trial of the Test of Memory Malingering (Tombaugh, 1996). Higher scores represent better performance on
the subtest.
b
Test score on the Structured Inventory of Malingered Symptomatology (Smith & Burger, 1997). Higher scores represent more malingered
symptomatology.
c
VAT-E subtest Immediate Recognition, Delayed Recognition, and Consistency are the underperformance measures of the VAT-E (Meyer & de Jonghe,
2017). Higher scores represent better performance on the subtest.

Table 3. The number of Korsakoff amnesia patients (n ¼ 20) and healthy comparisons (n ¼ 20) failing on the Test of Memory
Malingering (Trial 1 and 2), the Structured Inventory of Malingered Symptomatology (SIMS), and the subtests of the Visual
Association Test – Expanded.
Single tests Healthy Comparisons Korsakoff Amnesia
TOMM - Trial 1 3 (15%) 9 (45%) v2(1) ¼ 5.58, p ¼ .02, g ¼ .37
TOMM - Trial 2 0 (0%) 2 (10%) v2(1) ¼ 2.11, p ¼ .15, g ¼ .23
VAT-E Immediate Recognition 0 (0%) 1 (5%) v2(1) ¼ 1.02, p ¼ .31, g ¼ .16
VAT-E Delayed Recognition 0 (0%) 5 (25%) v2(1) ¼ 5.71, p ¼ .02, g ¼ .38
VAT-E Consistency Index 0 (0%) 5 (25%) v2(1) ¼ 5.71, p ¼ .02, g ¼ .38
SIMS 1 (5%) 0 (0%) v2(1) ¼ 1.02, p ¼ .31, g ¼ .16
Failing on both of the combined tests
VAT-E Profile analysis 0 (0%) 0 (0%)
TOMM – Trial 2 & Sims 0 (0%) 0 (0%)
VAT-E – Immediate Recognition & SIMS 0 (0%) 0 (0%)
TOMM – Trial 2 & VAT-E Immediate Recognition 0 (0%) 1 (5%) v2(1) ¼ 1.02, p ¼ .31, g ¼ .16
TOMM – Trial 2 & VAT-E Delayed Recognition 0 (0%) 2 (10%) v2(1) ¼ 2.11, p ¼ .15, g ¼ .23
TOMM – Trial 2 & VAT-E Consistency Index 0 (0%) 2 (10%) v2(1) ¼ 2.11, p ¼ .15, g ¼ .23
Note. IQ ¼ Intelligence Quotient based on the Dutch version of the National Adult Reading Test (Schmand et al., 1991); VAT-E ¼ Visual Association Test –
Extended (Meyer et al., 2017); TOMM ¼ Test of Memory Malingering (Tombaugh, 1996); SIMS ¼ Structured Inventory of Malingered Symptomatology
(Smith & Burger, 1997); IR ¼ Immediate Recognition; DR ¼ Delayed Recognition; N.A. ¼ Not Available because of ceiling scores on one of the indices.

Logistic regression suggesting reasonable to good ability to detect KA.

These results suggest that both the TOMM and VAT-
Linear regression collinearity diagnostics indicated strong
E subscales intended to index underperformance are
multicollinearity of the data structure for the indices of the
good in predicting whether participants were in the
TOMM, VAT-E, and SIMS, with a condition index of
amnesia patient or healthy comparison group.
>30 for four of the subscales. In a forward conditional
logistic regression analysis with all variables, the best
model fit was reached with only the VAT-E Delayed Proportional analysis
Recognition scale (v2(5) ¼ 24.55, p < .001), suggesting
Table 3 shows the proportions of patients and
good model fit, with Cox and Snell R2 ¼ .46, and a correct
healthy comparisons who failed the TOMM, the
classification of 82.5%. This minimal model suggests
VAT-E, and the SIMS, and who failed on two com-
strong interdependence of the PVT indices.
bined tests. As can be seen, 1–9 (5–45%) of the KA
patients failed the subtests on one or both PVTs.
Receiver Operating Characteristic (ROC)-analysis
The proportion of the two groups that failed on the
In ROC-analysis, the AUC of the SIMS is .61, suggest- TOMM – Trial 1 (v2(1) ¼ 5.58, p = .02, g ¼ .37), the
ing a low ability to predict KA. The AUC was .80 for VAT-E Delayed Recognition (v2(1) ¼ 5.71, p = .02,
Trial 1 of the TOMM, and .81 for Trial 2 of the g ¼ .38), and VAT-E Consistency scale (v2(1) ¼ 5.71,
TOMM, suggesting that both subtests of the TOMM p = .02, g ¼ .38), were statistically different between
were good in predicting KA. The AUC was .70 for the KA patients and controls. The results represented in
immediate recognition, .83 for the delayed recogni- Table 3 suggest that for both the TOMM and VAT-E,
tion, and .81 for the consistency score of the VAT-E, the number of false positives were higher in the KA
6 E. OUDMAN ET AL.

patients than in healthy comparisons. Combining civil litigants cases and forensic settings, false positives
PVTs was not effective in eliminating false positives are unacceptable (Rogers, 2008), while patients with
in KA patients, and had comparable false response severe cognitive deficits due to KA or other forms of
rates as the best out of two elements. Importantly, alcohol-related-brain-damage (ARBD) are relatively
performing profile analysis, or combining a PVT overrepresented in such populations (Ekstr€ om,
index with the SIMS was effective in eliminating Kristiansson, & Bj€ orksten, 2017). Therefore, the find-
false positives. ings in our study highlight the importance to interpret
PVT performance with caution.
In our study, a logistic regression approach with all
Discussion
the instruments generated a comparable level of pre-
The aim of our study was to identify whether patients diction as the VAT-E delayed recognition instrument.
with severe Korsakoff amnesia (KA) are able to pass Importantly, this finding highlights that the KA
Performance Validity Tests (PVTs) and Symptom patients that fail performance on the TOMM due to
Validity Tests (SVTs) despite the severity of their severe amnesia, also fail performance on the VAT-E.
amnesia. Moreover, we wanted to elucidate how the In Table 3 of our manuscript it is confirmed that
outcome of the TOMM, SIMS, and VAT-E could pre- combining PVTs was not fully successful in eliminat-
dict the presence of KA, and whether combining evi- ing the false positives. The most successful instrument
dence would reduce the effects of severe amnesia. Our was the profile analysis instrument of the VAT-E that
results show that PVT task performance was lower for combined scores from multiple resources. In line with
the TOMM and VAT-E in KA patients, and did lead Greher and Wodushek (2017), this finding highlights
to 5–45% false positive scores on the PVT subscales. the importance of profile analysis to support the find-
In logistic regression, Z-transformed indices could ings of PVTs, instead of relying solely on cutoff data.
predict KA with an accuracy of 82.5%, and the indi- A different approach that reduced sensitivity to severe
vidual PVT indices all had reasonable to good ability amnesia was to combine PVT indices with the SIMS.
to predict KA in ROC-analysis. Combining PVTs was It could be of practical value in populations with
not effective in reducing the number of false positives severe cognitive difficulties to investigate both under-
to zero, although profile analysis and including the performance and over-reporting of symptoms as two
SIMS as additional index were effective. Our results important factors of malingering.
suggest that the TOMM and VAT-E should be inter- Our study included a relatively new instrument to
preted with serious caution in patients with suspected assess both declarative memory functioning and per-
severe amnesia, since the severity of amnesia does formance validity, called the VAT-E. Earlier evidence
affect PVT performance. indicated that the VAT-E yields very promising results
The fundamental assumption of a PVT is insensi- in discriminating patients with Alzheimer’s disease,
tivity to cognitive dysfunction (Tombaugh, 1997). The Mild Cognitive Impairment (MCI), and healthy per-
results of the present study show that PVT perform- sons that feign memory deficits (Meyer et al., 2017).
ance is not insensitive to severe amnesia due to KA. In our study, not all KA patients were able to pass the
Earlier studies in dementia highlighted specificity rates VAT-E indices, although only one patient failed the
of 24% of the TOMM trial 2 in severe dementia VAT-E Immediate Recognition Trial. All patients
(Teichner & Wagner, 2004), up to 80% in mild to were able to pass the profile analysis score. A benefit
moderate dementia (Walter et al., 2014). Our findings of applying this novel instrument is that multiple indi-
indicate that specificity was 90% in KA patients on ces to detect underperformance can be combined to
the TOMM trial 2, but 75–95% on the indices of the more validly assess suboptimal performance. It would
VAT-E. Compared to the earlier reports in dementia, be of relevance to study the VAT-E in a larger sample
the KA patients performed somewhat better, although of amnesia patients to investigate whether the sensitiv-
compromised. Our findings contrast to the recent ity is comparable in a larger participant group.
observation by Erdodi and Rai (2017) that a single In a recent meta-analysis the SIMS was rated as a
error on the TOMM trial 2 already raises concerns good test that is well able to distinguish between
about the credibility of the patient, since 7 out of feigners and nonfeigners of complaints (van Impelen
20 KA patients (35%) had a single or more errors on et al., 2014). Earlier results also showed good sensitiv-
the second trial of the TOMM. A single error could ity and specificity in a sample of patients with cogni-
therefore also reflect cognitive disorders affecting the tive disorders (Smith & Burger, 1997). In the current
sensitivity of PVT test performance. Specifically in study it was found that patients with KA are able to
 APPLIED NEUROPSYCHOLOGY: ADULT 7

score sufficiently on the SIMS. None of the KA In conclusion, our study shows that patients with
patients showed exaggeration of symptoms. This severe amnesia are not always able to pass the
implies that the SIMS can be administered safely inpa- TOMM, and VAT-E. Since severity of the amnesia
tients that suffer from severe amnesia. In our sample, was directly related to performance on the PVTs, the
one healthy comparison reported more symptoms performance on those tests should be interpreted with
than the cutoff value. Since relatively lower IQ caution in patients with suspected severe amnesia.
patients were overrepresented in our sample, this Combining traditional PVTs was not maximally
could have caused this effect. This notion was earlier effective, but including profile analysis resulted in
made in the SIMS literature, although the relationship maximum insensitivity to severe amnesia. Also, com-
in the present study seems to be somewhat stronger, bining PVTs and the SIMS was effective in eliminat-
possibly highlighting an overrepresentation of lower ing false positives. In clinical use of PVTs, it is
IQ’s in the present sample. Future research should therefore relevant to combine levels of evidence to
investigate the SIMS in dementia, since no such stud- maximize the insensitivity to severe amnesia.
ies have been conducted earlier, but are particularly
relevant to validate the insensitivity of this test
for dementia. Disclosure statement
One could argue that the KA patients that under- No potential conflict of interest was reported by
performed in the present study were not false nega- the authors.
tives, but actual underperformers. Neuropsychological
tests are only able to measure behavior and not intent, Funding
which is essentially unknowable. We are therefore not
This work was not supported by any funding.
fully able to rule out the possibility that the KA
patients did underperform. However, patients did not
have incentives, such as litigations, to underperform ORCID
in our study, and were actively motivated by the test Erik Oudman [Link]
instructor to engage throughout the experiment.
Moreover, the patients in our study were not tested
with the PVTs and SVT as part of routine neuro-
psychological assessment, but were voluntarily partici- References
pating in this experiment, reducing the motivational American Psychiatric Association. (2013). Diagnostic and
effects. Moreover, the testing procedure was restricted statistical manual of mental disorders (DSM-5V R ).

to limit the effects of cognitive overload. Arlington, TX: American Psychiatric Association.
Bigler, E. D. (2012). Symptom validity testing, effort, and
Limitations of the present study include the gender
neuropsychological assessment. Journal of the
composition of the sample, which limits the generaliz-
International Neuropsychological Society, 18(04), 632–640.
ability of the findings to women. Earlier studies doi:10.1017/S1355617712000252
showed that men are overrepresented in KA popula- Bortnik, K. E., Horner, M. D., & Bachman, D. L. (2013).
tions, also reflected in the present study (Kopelman Performance on standard indexes of effort among
et al., 2009). A second limitation of this study is that patients with dementia. Applied Neuropsychology: Adult,
both patients and controls were not tested with a 20(4), 233–242. doi:10.1080/09084282.2012.695757
larger battery of neuropsychological tests, but rather Bush, S. S., Ruff, R. M., Tr€oster, A. I., Barth, J. T., Koffler,
S. P., Pliskin, N. H., … Silver, C. H. (2005). Symptom
performed only a few tests. One reason to restrict the
validity assessment: Practice issues and medical necessity:
number of tests in this study was the relative chance NAN Policy & Planning Committee. Archives of Clinical
for overloading the patients. Declarative memory tests Neuropsychology, 20, 419–426. doi:10.1016/
are particularly stressful for KA patients, because of [Link].2005.02.002
their generally low performance, and lack of insight Dandachi-FitzGerald, B., Ponds, R. W. H. M., Peters,
into their test performance (Walvoort, van der M. J. V., & Merckelbach, H. (2011). Cognitive underper-
Heijden, Kessels, & Egger, 2016). Moreover, the sam- formance and symptom over-reporting in a mixed psy-
chiatric sample. The Clinical Neuropsychologist, 25(5),
ples in the present study were relatively low and,
812–828. doi:10.1080/13854046.2011.583280
therefore, restricted the statistical power of some of Dandachi-FitzGerald, B., Ponds, R. W., & Merten, T.
the models. This fact could be explained by the rela- (2013). Symptom validity and neuropsychological assess-
tive scarcity of detoxified KA patients available for ment: A survey of practices and beliefs of neuropsycholo-
neuropsychological testing. gists in six European countries. Archives of Clinical
8 E. OUDMAN ET AL.

Neuropsychology, 28(8), 771–783. doi:10.1093/arclin/ Meyer, S., & de Jonghe, J. (2017). Visuele associatietest -
act073 extended: Handleiding. Amsterdam, Netherlands: Hogrefe.
Dorociak, K. E., Schulze, E. T., Piper, L. E., Molokie, R. E., Meyer, S. R., de Jonghe, J. F., Schmand, B., & Ponds, R. W.
& Janecek, J. K. (2018). Performance validity testing in a (2017). The Visual Association Test-Extended: A cross-
clinical sample of adults with sickle cell disease. The sectional study of the performance validity measures. The
Clinical Neuropsychologist, 32(1), 81–97. doi:10.1080/ Clinical Neuropsychologist, 31(4), 798–813. doi:10.1080/
13854046.2017.1339830 13854046.2017.1280181
Ekstr€om, A., Kristiansson, M., & Bj€ orksten, K. S. (2017). Mittenberg, W., Patton, C., Canyock, E. M., & Condit,
Dementia and cognitive disorder identified at a forensic D. C. (2002). Base rates of malingering and symptom
psychiatric examination - a study from Sweden. BMC exaggeration. Journal of Clinical and Experimental
Geriatrics, 18, 219. Neuropsychology, 24(8), 1094–1102. doi:10.1076/
Erdodi, L. A., & Rai, J. K. (2017). A single error is one too jcen.24.8.1094.8379
many: Examining alternative cutoffs on Trial 2 of the Rogers, R. (Ed.). (2008). Clinical assessment of malingering
TOMM. Brain Injury, 31(10), 1362–1368. doi:10.1080/ and deception. New York, USA: Guilford Press.
02699052.2017.1332386 Roth, M., Tym, E., Mountjoy, C. Q., Huppert, F. A.,
Folstein, M., Folstein, S. E., & McHugh, P. R. (1975). Hendrie, H., Verma, S., & Goddard, R. (1986).
“Mini-Mental State” a practical method for grading the CAMDEX. A standardised instrument for the diagnosis
cognitive state of patients for the clinician. Journal of of mental disorder in the elderly with special reference to
Psychiatric Research, 12(3), 189–198. doi:10.1016/0022- the early detection of dementia. British Journal of
3956(75)90026-6 Psychiatry, 149(6), 698–707.
Greher, M. R., & Wodushek, T. R. (2017). Performance Rudman, N., Oyebode, J. R., Jones, C. A., & Bentham, P.
validity testing in neuropsychology: Scientific basis and (2011). An investigation into the validity of effort tests in
clinical application—A brief review. Journal of Psychiatric a working age dementia population. Aging & Mental
Practice, 23(2), 134–140. doi:10.1097/PRA.00000000 Health, 15(1), 47–57. doi:10.1080/13607863.2010.508770
00000218 Schmand, B. A., Bakker, D., Saan, R. J., & Louman, J.
Hampson, N. E., Kemp, S., Coughlan, A. K., Moulin, (1991). De Nederlandse Leestest voor Volwassenen: een
C. J. A., & Bhakta, B. B. (2014). Effort test performance maat voor het premorbide intelligentieniveau [The Dutch
in clinical acute brain injury, community brain injury,
Adult Reading Test: A measure of premorbid intelli-
and epilepsy populations. Applied Neuropsychology: Adult,
gence]. Tijdschrift Voor Gerontologie en Geriatrie, 22,
21(3), 183–194. doi:10.1080/09084282.2013.787425
15–19.
Iverson, G. L. (2011). Test of memory malingering.
Schmand, B., Walstra, G., Lindeboom, J., Teunisse, S., &
Encyclopedia of Clinical Neuropsychology (2494-2496).
Jonker, C. (2000). Early detection of Alzheimer’s disease
New York, USA: Springer.
using the Cambridge Cognitive Examination
Kopelman, M. D. (2002). Disorders of memory. Brain : A
(CAMCOG). Psychological Medicine, 30(3), 619–627. doi:
Journal of Neurology, 125(Pt 10), 2152–2190.
10.1017/S0033291799002068
Kopelman, M. D., Thomson, A. D., Guerrini, I., &
Slick, D. J., Tan, J. E., Strauss, E., Mateer, C. A., Harnadek,
Marshall, E. J. (2009). The Korsakoff syndrome: Clinical
aspects, psychology and treatment. Alcohol & Alcoholism, M., & Sherman, E. M. (2003). Victoria Symptom Validity
44, 148–154. doi:10.1093/alcalc/agn118 Test scores of patients with profound memory impair-
Lindeboom, J., & Matto, D. (1994). Digit Span and Knoxx ment: Nonlitigant case studies. The Clinical
Blocks as concentration tests for elderly [in Dutch]. Neuropsychologist, 17(3), 390–394. doi:10.1076/
Tijdschrift Voor Gerontologie en Geriatrie, 25, 63–68. clin.17.3.390.18090
Lindeboom, J., Schmand, B., & Christensen, N. (2003). Sweet, J. J., Benson, L. M., Nelson, N. W. & Moberg,
Visual Association Test: VAT; Handleiding. Leiden, (2015). The American Academy of Clinical
Netherlands: PITS. Neuropsychology, National Academy of
Liu, Z., Dong, J., Zhao, X., Chen, X., Lippa, S. M., Caroselli, Neuropsychology, and Society for Clinical
J. S., & Fang, X. (2016). Assessment of feigned cognitive Neuropsychology (APA Division 40) 2015 TCN
impairment in severe traumatic brain injury patients with Professional Practice and ‘Salary Survey’: Professional
the Forced-choice Graphics Memory Test. Brain and Practices, Beliefs, and Incomes of U.S.
Behavior, 6(12), e00593. doi:10.1002/brb3.593 Neuropsychologists. The Clinical Neuropsychologist, 29,
Loring, D. W., Larrabee, G. J., Lee, G. P., & Meador, K. J. 1069–1162. doi:10.1080/13854046.2016.1140228
(2007). Victoria Symptom Validity Test performance in a Smith, G. P., & Burger, G. K. (1997). Detection of malinger-
heterogenous clinical sample. Clinical Neuropsychologist, ing: Validation of the Structured Inventory of Malingered
21(3), 522–531. doi:10.1080/13854040600611384 Symptomatology (SIMS). Journal of the American
Martin, P. K., Schroeder, R. W., & Odland, A. P. (2015). Academy of Psychiatry and the Law Online, 25, 183–189.
Neuropsychologists’ validity testing beliefs and practices: Teichner, G., & Wagner, M. T. (2004). The Test of Memory
A survey of North American professionals. The Clinical Malingering (TOMM): Normative data from cognitively
Neuropsychologist, 29(6), 741–776. doi:10.1080/ intact, cognitively impaired, and elderly patients with
13854046.2015.1087597 dementia. Archives of Clinical Neuropsychology, 19(3),
McGuire, C., Crawford, S., & Evans, J. J. (2019). Effort test- 455–464. doi:10.1016/S0887-6177(03)00078-7
ing in dementia assessment: A systematic review. Tombaugh, T. N. (1996). Test of memory malingering:
Archives of Clinical Neuropsychology, 34(1):114–131. TOMM. New York, USA/Toronto, Canada: MHS.
 APPLIED NEUROPSYCHOLOGY: ADULT 9

Tombaugh, T. N. (1997). The Test of Memory Malingering validity measure. Clinical Neuropsychologist, 28(7),
(TOMM): Normative data from cognitively intact and 1197–1208. doi:10.1080/13854046.2014.960454
cognitively impaired individuals. Psychological Walvoort, S. J. W., van der Heijden, P. T., Kessels, R. P. C.,
Assessment, 9(3), 260. doi:10.1037/1040-3590.9.3.260 & Egger, J. I. M. (2016). Measuring illness insight in
Van Impelen, A., Merckelbach, H., Jelicic, M., & Merten, T. patients with alcohol-related cognitive dysfunction
(2014). The Structured Inventory of Malingered using the Q8 questionnaire: A validation study.
Symptomatology (SIMS): A systematic review and meta- Neuropsychiatric Disease and Treatment, 12, 1609–1615.
Wilson, B. A., Alderman, N., Burguess, P. W., Emslie, H., &
analysis. The Clinical Neuropsychologist, 28(8),
Evans, J. J. (1996). In: Behavioural Assessment of the
1336–1365. doi:10.1080/13854046.2014.984763 Dysexecutive Syndrome (BADS). Souza Ricardo, O., Schmidt
Moerman - van den Brink, W. G., van Aken, L., Verschuur, Sergio L., translators. Bury St Edmunds, U.K.: Thames
E. M. L., Walvoort, S. J. W., Egger, J. I. M., & Kessels, Valley Test Company; Rio de Janeiro, Brazil: Cogniç~ao.
R. P. C. (2019). Executive dysfunction in patients with Wodushek, T. R., & Greher, M. R. (2017). Performance val-
Korsakoff’s syndrome: A theory-driven approach. Alcohol idity testing in neuropsychology: Methods for measure-
and Alcoholism, 54(1):23–29. doi:10.1093/alcalc/agy078. ment development and maximizing diagnostic accuracy.
Walter, J., Morris, J., Swier-Vosnos, A., & Pliskin, N. Journal of Psychiatric Practice, 23(3), 214–220. doi:
(2014). Effects of severity of dementia on a symptom 10.1097/PRA.0000000000000233