Table 2 Review of the literature of set-shifting, central coherence and decision-making in individuals recovered from anorexia nervosa. The articles are presented chronologically according to year of publication

Tchanturia et al., 2002 [10]

Set-shifting

Fixed set task

To examine perceptual and cognitive set-shifting

AN = 30

AN = 25.2 (6.7)

Significant differences between AN/AN-REC compared to HC

Impaired set-shifting could represent a vulnerability factor for AN

Cognitive shift task

AN-REC = 16

AN-REC = 30.0 (6.0)

HC = 23

HC = 27.6 (6.4)

Tchanturia et al., 2004 [15]

Set-shifting

TMT

To investigate set-shifting in current and past AN patients

AN = 34

AN = 27.2 (8.3)

Set-shifting difficulties was observed in the AN group, and to a lesser degree in the AN-REC group

Some aspects of set-shifting in AN appear to be a trait rather than state marker

Brixton test

AN-REC = 18

AN-REC = 28.4 (6.8)

Picture set test

HC = 36

Cat bat tests

HC = 25.9 (4.8)

Uznadze illusion task

Holliday et al., 2005 [16]

Set-shifting

Haptic illusion task

To investigate whether set-shifting difficulties are familial

47 pairs of sisters disconcordant for AN: AN-REC = 23

AN (AN-REC + AN+ AN-WR) = 26.3 (10.2)

Set-shifting difficulties persists after recovery

Suggest that set-shifting difficulties are trait characteristics and may inform the search for the endophenotype of AN

Brixton Test

TMT

CatBat task

HS = 27.6 (9.6)

AN = 21

AN-WR = 3

HC = 26.5 (6.1)

HS = 47

HC = 47

Tchanturia et al., 2007 [17]

Decision making

IGT

To determine whether decision-making is impaired in AN-REC

AN = 29

AN = 28.5 (9.17)

No differences between AN-REC and HC. AN grop did poorer than AN-REC and HC

Impaired IGT performance could be a consequence of starvation

HC = 29

HC = 26.3 (7.9)

AN-REC = 14

AN-REC = 28.9 (7.4)

Lopez et al., 2009 [19]

Central coherence

Embedded figures test Unsegmented/segmented block design

To examine whether AN-REC women have weak central coherence

AN-REC = 42

AN-REC = 25

Weak central coherence in the AN-REC group

Suggest that weak central coherence is a stable trait and an endophenotype for AN

HC = 42

HC = 26

RCFT

SCT

Homograph reading task

Nakazato et al., 2009 [18]

Set-shifting

WCST

To establish whether set-shifting difficulties are present in AN and AN-REC

AN = 29

AN = 28.3 (11.0)

No significant differences between AN-REC and HC

Strongly suggest that impairment in set-shifting normalizes with recovery

AN-REC = 18

AN-REC = 32.2 (11.1)

HC = 28

HC = 26.9 (5.8)

Nakazato et al., 2010 [20]

Set-shifting

WCST

TMT

Determine whether serum glutamine is associated with set-shifting ability

AN = 27

AN = 27.7 (10.6)

No differences between AN-REC and HC

Serum concentrations does not appear to be associated with executive functions

AN-REC = 32.2 (11.1)

HC = 26.9 (5.8)

AN-REC = 18

HC = 28

Roberts et al., 2010 [21]

Set-shifting

TMT

To add clarity to set-shifting in EDs

AN-R = 35

AN-R = 23.71 (6.39)

Some evidence for poorer set shifting in AN – REC

It is likely a familial trait, and related to the maintenance of the illness

WCST

AN-BP = 33

Brixton test

BN = 30

AN-BP = 25.58 (7.64)

Haptic illusion

AN-REC = 30

BN = 26.43 (6.84)

AN sister = 30

BN sister =20

AN-REC = 32.13 (11.64)

HC = 88

AN sister = 24.23 (6.44)

BN sister = 27.60 (8.71)

HC = 28.43 (8.47)

Tenconi et al., 2010 [14]

Set-shifting Central coherence

WCST

To explore suitability for endophenotypes in ED

AN = 60

AN = 26.2 (6.9)

No differences between AN and AN-REC

Impaired set-shifting and low central coherence might be an endophenotype of AN

TMT

AN-WR = 63

HC = 27.4 (4.5)

RCFT

AN-REC = 30

HC = 120

Harrison et al., 2011 [13]

Central coherence

RCFT

GEFT

FPT

To replicate findings of weak central coherence

AN = 50

ED group = 27.13 (9.3)

Superior detail processing skills was associated with having an ED and recovered from AN

AN-REC are skilled at detailed processing. Weak central coherence may be a factor that perpetuates ED behaviors

BN = 48

AN-REC = 35

AN-REC = 29.00 (10.62)

HC = 89

HC = 28.5 (9.93)

Bühren et al., 2012 [12]

Set-shifting

Visual set-shifting task

To investigate set-shifting in AN before and after weight gain

AN = 28

AN = 15.6 (1.5)

No deficits in set-shifting abilities.

Speculate that findings could be related to short duration of illness

HC = 27

HC = 15.0 (1.7)

Danner et al., 2012 [22]

Set-shifting Central coherence Decision making

BCST

To examine set-shifting in women AN-REC

AN = 16

AN = 25.63 (5.41)

Poor set-shifting and decision making in AN-REC compared to HC. No differences between the groups in central coherence

Suggest that impaired decision-making and set-shifting is a stable trait in AN

RCFT

AN-REC = 15

IGT

HC = 15

AN-REC = 24.33 (4.72)

HC = 25.8 (4.69)

Favaro et al., 2012

Central coherence

RCFT

To explore functional connectivity of networks involved in visuospatial and somatosensory processing

AN = 29

AN = 25.8 (6.9)

No differences between AN-REC and HC in central coherence

Results could be due to a milder form of the illness in the AN-REC group, measures by minimum BMI and duration of illness

AN-REC = 16

AN-REC = 23.8 (4.8)

HC = 26

HC = 26.7 (6.7)

Harrison et al., 2012 [24]

Set-shifting

WCST

To explore cognitive and social emotional functioning

ED = 100

ED-REC = 35

HC = 90

N/A

ED-REC did not differ from the acute group.

Suggest that cognitive style is a trait.

Central coherence

Brixton task

FPT

RCFT

Lindner et al., 2012 [25]

Decision making

IGT

To examine decision-making and planning in AN-REC

AN-REC = 100

AN-REC = 34.49 (7.13)

AN-REC did better in decision-making

Findings are in contrast to previous findings of impaired decision making in AN-REC

HC = 100

HC = 34.53 (7.26)

Tchanturia et al., 2012 [26]

Set-shifting

WCST

To explore WCST performance and other clinical outcomes

AN = 171

AN = 25.4 (8.2)

AN-REC showed better performance than AN, but did more perseverative errors than HC

This large dataset supports previous studies which indicate poor cognitive flexibility in people with EDs.

BN = 82

BN = 27.3 (8.3)

AN-REC = 90

AN-REC = 30.7 (11.1)

HC = 199

HC = 27.7 (8.8)

Lindner et al., 2013 [27]

Central coherence

RCFT

To examine central coherence as a possible endophenotype in AN

AN-REC = 100

AN-REC = 34.49 (7.13)

AN-REC group showed better accuracy in the copy condition

No inefficiencies in global processing but a superior local processing

HC = 100

HC = 34.53 (7.26)

Lindner et al., 2014 [28]

Set-shifting

BCST

To explore whether set-shifting is inefficient after full recovery of AN

AN-REC = 100

AN-REC = 34.49 (7.13)

AN-REC achieved fewer categories, more perseverations and spent less time for shifting set

Suggest that set-shifting is inefficient after full recovery

HC = 100

HC = 34.53 (7.26)

Ritschel et al., 2015 [29]

Decision making

Intertemporal choice task

To investigate delay discounting in ill and recovered AN

AN = 34

AN = 15.29 (2.7)

No group differences in delay discounting

Suggest that delay discounting is not a trait marker for AN

AN-REC = 33

AN-REC = 21.67 (3.1)

HC = 54

HC = 18.75 (4.4)

Talbot et al., 2015 [9]

Set-shifting and central coherence

WCST

Matching familiar figures test

RCFT

To investigate whether impaired set-shifting and weak central coherence represent state or trait

AN = 24

AN = 21.0 (18-

Poorer set-

This study found no

AN-WR = 10

27)^a

shifting in AN –REC and AN-WR compared to HC.

support that weak central coherence is an endophenotype for AN

AN-REC = 15

AN-WR = 21.5 (19–25) ^a

HC = 43

AN-REC = 24.0 (21–32) ^a

No differences between the groups on measures of local and global processing

HC = 19.0 (18–25) ^a

Ely et al., 2016 [30]

Set shifting

Color-Word interference test

To identify deficits specific to inhibition or task-switching

AN-REC = 47

AN-REC = 26.68 (1.83)

Differences between the groups trended toward – but was not – significant

Suggest that cognitive control impairments in AN is related to anxiety, and not a neuropsychological deficit

HC = 24

HC = 25.08 (6.11)

Sultson et al., 2016 [31]

Set-shifting

BCST

To investigate whether activation in frontal and parietal regions is associated with set-shifting ability

AN = 16

AN = 25.57 (5.8)

No differences between AN-REC and HC, but a trend towards more perseverative errors and completing less categories than HC

Higher activation in frontal regions involved in self-referential processing and cognitive control is associated with poor set-shifting ability in AN-REC

AN-REC = 15

AN-REC = 24.79 (4.5)

HC = 15

HC = 25.80 (4.7)

Bentz et al., 2017 [11]

Set shifting and local processing

TMT

To investigate impairments of social functioning and potential associations with neurocognitive functions

AN = 43

AN = 16.1 (1.5)

No differences between the groups on neurocognitive functions

Young AN and AN-REC did not differ from HC

GEFT

AN-REC = 28

AN-REC = 18.4 (1.6)

HC = 41

HC = 17.7 (2.2)