If you don't remember your password, you can reset it by entering your email address and clicking the Reset Password button. You will then receive an email that contains a secure link for resetting your password
If the address matches a valid account an email will be sent to __email__ with instructions for resetting your password
Department of Psychiatry and Clinical Psychobiology, University of Barcelona, SpainCentro de Investigación Biomédica en Red de Salud Mental, Barcelona, SpainInstitut Clínic de Neurociènces, Hospital Clinic, Barcelona, Spain
Department of Psychiatry and Clinical Psychobiology, University of Barcelona, SpainInstitut d’Investigacions Biomèdiques August Pi i Sunyer, Barcelona, Spain
Department of Psychiatry and Clinical Psychobiology, University of Barcelona, SpainCentro de Investigación Biomédica en Red de Salud Mental, Barcelona, SpainInstitut Clínic de Neurociènces, Hospital Clinic, Barcelona, Spain
Department of Psychiatry and Clinical Psychobiology, University of Barcelona, SpainCentro de Investigación Biomédica en Red de Salud Mental, Barcelona, SpainInstitut Clínic de Neurociènces, Hospital Clinic, Barcelona, Spain
Department of Psychiatry and Clinical Psychobiology, University of Barcelona, SpainCentro de Investigación Biomédica en Red de Salud Mental, Barcelona, SpainInstitut Clínic de Neurociènces, Hospital Clinic, Barcelona, Spain
Department of Psychiatry and Clinical Psychobiology, University of Barcelona, SpainInstitut d’Investigacions Biomèdiques August Pi i Sunyer, Barcelona, Spain
Cognitive remediation therapy positively affects cognition and daily functioning in patients with schizophrenia. However, studies on the underlying neurobiological mechanisms of this treatment are scarce. The aim of the current study was to investigate functional and structural connectivity brain changes in schizophrenia patients after cognitive remediation therapy using a whole-brain approach that combined functional magnetic resonance imaging and diffusion tensor imaging.
Methods
A randomized controlled trial with 30 schizophrenia outpatients and 15 healthy volunteers. A strategy-learning-based treatment was used as a cognitive remediation therapy. A social skills training that provides useful information about illness management was used as an active control. We investigated changes in the pattern of functional connectivity assessed during an n-back task by tensorial independent component analysis as implemented in the multivariate exploratory linear decomposition into independent components and in the fractional anisotropy index of white matter integrity using tract-based spatial statistics.
Results
Brain networks activation pattern significantly changed in patients exposed to the cognitive treatment in the sense of normalizing toward the patterns observed in healthy control subjects. Additionally, in white matter, they showed an increase in fractional anisotropy index in the anterior part of the genu of the corpus callosum. Cognitive improvement, functional, and also structural changes showed statistically significant correlations.
Conclusions
Improvement in brain functioning detected after cognitive remediation therapy in schizophrenia patients might be based on an increase of the interhemispheric information transfer between the bilateral prefrontal cortexes via the corpus callosum.
Cognitive remediation therapy (CRT) is an evidence-based treatment that seems to improve neurocognition and positively affect daily functioning in patients with schizophrenia (
What are we changing with neurocognitive rehabilitation? Illustrations from two single cases of changes in neuropsychological performance and brain systems as measured by SPECT.
) reported how two patients showed changes in frontal perfusion patterns after the cognitive treatment measured through single photon emission computed tomography (SPECT) procedures. Penadés et al. (
) also found an increase in prefrontal blood flow during task performance following cognitive treatment in a case report study. These results were replicated in a small sample of eight patients (
Preliminary evidence of improved verbal working memory performance and normalization of task-related frontal lobe activation in schizophrenia following cognitive exercises.
) found that verbal memory improvement after training was associated with increased task-related activation in the same brain region that was activated during memory tasks in healthy individuals. Unfortunately, these initial studies had small samples and lack of real placebo comparison groups. Wykes et al. (
) conducted the first randomized, controlled study in which 12 patients were randomly assigned to control therapy or CRT. Only the group receiving CRT significantly increased their brain activation in regions associated with working memory, particularly the inferior frontal gyrus. Recently, Haut et al. (
) conducted a quasi-randomized study with nine patients receiving CRT, nine patients receiving control therapy (social skills training), and nine healthy control (HC) subjects. Those patients receiving CRT showed increased activation in the dorsolateral prefrontal cortex, anterior cingulate, and frontopolar cortex. Finally, Bor et al. (
) explored the impact of CRT in 17 patients and 15 healthy volunteers with fMRI in a randomized controlled trial. Following treatment, the CRT group exhibited an increased ability to activate the prefrontal regions that could be subserving attention and working memory systems.
Although these studies led the way in showing the possibility of concomitant changes in brain activity with CRT, they did not take into account the networks of activation involved in cognitive functioning. Thus, hypofrontality has been supported by meta-analysis (
). Interestingly, not only hypoactivation but also compensatory hyperactivation from other nonspecific prefrontal cortex areas such as the ventral regions have been described (
). Notwithstanding, cognitive impairment in schizophrenia seems to be the consequence of widespread alterations of the connectivity between various brain networks and is not restricted to the dorsal prefrontal cortex (
). It is the so-called default mode network (DMN) that involves different interconnected regions that should be highly active at rest but that should be deactivated during performance of cognitive tasks (
Furthermore, only one study has tested the effects of CRT with structural neuroimaging methods showing compelling results. After a 2-year follow-up, Eack et al. (
Neuroprotective effects of cognitive enhancement therapy against gray matter loss in early schizophrenia: Results from a 2-year randomized controlled trial.
) were able to demonstrate that patients following cognitive treatment showed better preservation of gray matter volumes in regions that are involved in cognitive impairment than the patients who did not. Despite these encouraging results, to date no study of CRT has been conducted to explore the impact of the CRT on structural connectivity that integrates structural and functional MRI techniques. Thus, we hypothesized that the use of a strategy-based CRT would improve efficiency in different functional networks and that it also would produce some changes in structural connectivity.
Methods and Materials
A controlled, randomized study was carried out with three groups: patients receiving cognitive treatment, patients receiving a different psychological intervention as an active control, and a HC group. Randomization was independently conducted by author C.G. He took no part in the implementation of assignments, the generation of the allocation sequence being his only role. On recruitment, patients were randomly assigned to either CRT or social skills training (SST) using computer-generated random numbers and treated for 4 months in their respective treatment condition. All were then assessed 2 to 3 days before the first treatment session and 2 to 3 days after the last session through neuropsychological battery and MRI. To identify functional and structural global changes related to CRT, a longitudinal analysis of patients and HC subjects pre- and posttreatment MRI data was carried out. In addition, preprocessing and analysis of cross-sectional fMRI data at baseline were performed using the same protocol. Researchers involved in those analyses were completely blinded to the treatment assignation.
Participants
Patients (n = 35) were recruited by the Schizophrenia Unit at the Hospital Clinic, which serves part of the Barcelona area. The inclusion criteria were 1) age less than 55 years, 2) diagnosis of schizophrenia confirmed following the Structured Clinical Interview for DSM-IV Axis I Disorders (
First MB, Spitzer RL, Gibbon M, Williams JB (1997): Structured Clinical Interview for DSM-IV Axis I Disorders—Clinician Version (SCID-CV) [Spanish Edition: Barcelona: Masson SA, 2001]. Washington, DC: American Psychiatric Press
), and 4) presence of cognitive impairments confirmed by a battery of neuropsychological tests. Exclusion criteria were 1) Vocabulary test of the Wechsler Adult Intelligence Scale—Third Edition below 4 (scaled score), 2) organic cerebral diseases or primary diagnosis of substance abuse, 3) psychotic exacerbation in the previous 6 months, and 4) plan to change medication during the treatment phase. Healthy control participants (n = 15) were screened for the presence of lifetime Axis I psychotic or mood disorders using the Structured Clinical Interview for DSM-IV (nonpatient version) and for the presence of a first-degree relative with schizophrenia. Following a complete description of the study to participants, written informed consent was obtained. There were no significant differences between groups for patients, but HC subjects performed significantly better in cognitive variables (Table 1).
Table 1Demographic Clinical and Cognitive Variables
Some patients were polymedicated, increasing the sample size for medication. Five patients were taking two antipsychotics, three of whom were from the CRT and 2 from the SST.
a Some patients were polymedicated, increasing the sample size for medication. Five patients were taking two antipsychotics, three of whom were from the CRT and 2 from the SST.
) and is used to facilitate the strategy-learning in tasks of progressive complexity adopting an errorless learning approach. As much as possible, each task was set at the subject’s own pace with scaffolding as the main instructional technique. This involves an instructor extending a learner’s ability by providing support in those aspects of a task that the learner cannot accomplish, while removing assistance in those areas where competence has been achieved. The patients received 40 sessions: 1-hour sessions two or three times a week over 4 months. The Cognitive Shift Module aims to address flexibility in thinking and information-set maintenance, both of which presumably require the capacity to effectively engage and disengage activated neural network processing. The Working Memory Module aims to target the executive processes central to memory control and has patients work with as many as two to five information sets at a time. The primary target of the Planning Module is self-ordered, goal-oriented, set/schema formation and manipulation—that is, the application of the practiced processes, such as working memory, to tasks requiring planning.
SST. Patients in the treatment control group received 40 hours of treatment based on a behaviorally oriented SST. The intervention was adapted from the manualized therapy protocol Symptom Management Module from the University of California—Los Angles skills training modules (
) but they are not expected to have any effects on neurocognition. The module provides information about schizophrenia and its treatment and trains the participants to identify and monitor warning signs of relapse, to manage warning signs by developing a crisis plan, to cope with persistent symptoms, and to avoid and refuse alcohol and nonmedication drugs. The role of the therapist was to inform the participant of the relevant skills and their benefits, to enhance skill acquisition by modeling and role-play exercises, to anticipate outcome problems that could interfere with the use of the skill, and to encourage the transfer of targeted skills into real-world settings by in vivo exercises and homework assignments.
MRI Scanning Protocol
Functional and structural imaging data were acquired on a 3-T MRI scanner (Magnetom Trio Tim, Siemens Medical Systems, Germany). During fMRI a set of T2-weighted 280 volumes were acquired (repetition time [TR] = 2000 msec, echo time [TE] = 29 msec, slice thickness = 3 mm, distance factor = 25%, field of view [FOV] = 240 mm, matrix size = 128×128) providing whole brain coverage. A T1-weighted structural image was also acquired for each subject with magnetization prepared rapid acquisition gradient-echo three-dimensional protocol (TR = 2300 msec, TE = 2.98 msec, 240 slices, slice thickness = 1 mm, FOV = 256 mm, matrix size = 256 × 256). Finally, diffusion tensor imaging (DTI) was conducted in the same scanning session with single shot diffusion weighted echo-planar imaging in the axial plane with diffusion sensitization gradients applied in 60 noncolinear directions with a b value of 0 and 1000 s/mm2 (TR = 9300 msec, TE = 94 msec, 65 slices, slice thickness = 2 mm, FOV= 240, matrix size= 122×122). Image analysis was performed using Oxford Centre Functional MRI of the Brain software library (FSL version 4.1; http://www.fmrib.ox.ac.uk/fsl).
fMRI Data
Experimental Setup and N-back Paradigm. The task, adapted from Callicott et al. (
), involved monitoring locations of dots (presentation time: 450 msec; interstimulus-interval: 500 msec) within a diamond-shaped box at a given delay from the original occurrence (0-back, or 2-back; Figure 1). During the 0-back test, participants identified the number currently seen. In the 2-back load, participants encoded the current stimulus while indicating the number seen two stimuli previously. Participants responded by pressing the appropriate button on a four-button box, which has the same configuration and position as the diamond. Short resting periods (in which a white cross was shown on a black screen) were introduced after each 2-back condition. Every block consisted of 18 trials (6×0-back, 6×2-back, and 6×rest) each of 90-second duration. The sequence consisted of a total of six blocks. Stimuli were presented inside the 3T magnetic resonance machine with Visual Stim Digital MRI Compatible High Resolution Stereo three-dimensional glasses (Resonance Technology, Northridge, California) and Presentation version 10.1 (Neurobehavioral Systems, Albany, California) running on Windows XP (Microsoft, Redmond, Washington).
), spatial smoothing using a Gaussian kernel of 5 mm, mean-based intensity normalization of all volumes by the same factor (four-dimensional grand mean scaling), high pass temporal filtering (90 sec), and Gaussian low-pass temporal filtering (full width at half maximum = 2.8). The functional scans were registered to the Montreal Neurological Institute 152 standard space by using affine registration with FLIRT (
After this preprocessing, fMRI analysis was carried out using tensorial independent component analysis (TICA) as implemented in the Multivariate Exploratory Linear Decomposition into Independent Components (MELODIC Version 3.1) (
) tool, part of FSL software. TICA is a data-driven approach that decomposes the data into a set of independent components (ICs). Model order was estimated using the Laplace approximation to the Bayesian evidence for a probabilistic principal component analysis model (
). Estimated component maps were divided by the standard deviation of the residual noise and thresholded by fitting a mixture model to the histogram of intensity values. Z (Gaussianized t/F) statistic images were thresholded using clusters determined by Z≥2.3 and a cluster significance threshold of p≤.05 corrected for multiple comparisons.
DTI Preprocessing and Analysis. DTI preprocessing and analysis was performed using tools from FSL (
). Image artifacts due to eddy current distortions were minimized by registering the diffusion images to the b0 images. The registered images were skull stripped using BET (
). After calculation of the FA map for each subject, we implemented a voxelwise statistical analysis of the FA data using Tract-Based Spatial Statistics v1.2 (TBSS) (
). TBSS uses a carefully tuned nonlinear registration followed by the projection of the nearest maximum FA values onto a skeleton derived from a mean FA image to improve the sensitivity of the DTI analysis.
Statistical Analysis
For the neuropsychological outcomes an intent-to-treat analyses were first conducted with all participants who started the treatment sessions. A repeated-measures analysis of variance (ANOVA) design was used to test differences in cognitive variables between conditions at baseline and endpoint. Next, analyses of covariance comparing the two groups at baseline and endpoint assessments, including baseline as a covariate, were also performed.
Regarding neuroimaging data, a longitudinal design included the three groups at two time points with the 30 patients (15 CRT and 15 SST) who finished the neuroimaging protocol and the 15 HC subjects. Longitudinal data were analyzed with a repeated-measures ANOVA, and cross-sectional fMRI data were analyzed with one-way ANOVA, using permutation-based nonparametric inference within the framework of the general linear model (5000 permutations) (
) as implemented in the FSL randomize tool. Results were considered significant at p < .05, family-wise error (FWE)-corrected, using threshold-free cluster (TFCE), a method which avoids using an arbitrary threshold for the initial cluster-formation (
). Finally, motion parameters were extracted for every subject calculating the mean of displacement to establish a head movement correction. Then, a one-factor ANOVA was applied to test between-group differences. Finally, Pearson’s coefficients were calculated to determine the relationship between MRI changes and neuropsychological improvements.
Results
Neuropsychological Outcomes
The result of the overall intent-to-treat analysis with the addition of baseline covariates was significant for time by condition (F2,47 = 3.918; p = .029; effect size .175) and also for different cognitive measures (Table 2). Additionally, univariate analyses showed a beneficial effect of CRT on different cognitive tests and cognitive domains (Figure 2), particularly for executive function (F2,47 = 8.469; p = .001; effect size .314), Verbal Memory (F2,47 = 6.029; p = .005; effect size .236) and Nonverbal Memory (F2,47 = 10.418; p<.001; effect size .348). Changes on working memory and psychomotor speed were not significant. More to the point, performances during scanning in the n-back task were similar at baseline and the endpoint for the three groups with no change after treatment.
Table 2Neuropsychological Outcomes
Pre Mean (SD)
Baseline Between Groups
Baseline CRT vs. SST
Post Mean (SD)
Time×Group
Time×Group CRT vs. SST
0-Back Correct Hits
CRT
70.8 (.4)
F2,42 = 1.875
F1,29 = 1.875
70.6 (.1)
F2,42 = 1.875
—
SST
70.7 (.4)
p = .247
p = .247
70.8 (.4)
p = .247
HC
71.0 (1.5)
70.8 (.3)
2-Back Correct Hits
CRT
39.2 (6.2)
F2,42 = 1.709
F1,29 = 1.709
42.2 (5.2)
F2,42 = 1.433
—
SST
43.5 (1.2)
p = .227
p = .249
48.5 (2.1)
p = .227
HC
43.1 (1.9)
47.3 (6.8)
WAIS-III, Digit Span
CRT
12.53 (9.77)
F2,47 = 9.089
F1,34 = 2.720
14.29 (3.06)
F2,47 = .653
—
SST
14.38 (2.91)
p<.001
p = .109
16.69 (11.8)
p = .216
HC
18.01 (4.63)
16.13 (7.71)
WAIS-III, Arithmetic
CRT
7.82 (2.16)
F2,47 = 33.47
F1,34 = 5.351
11.24 (7.12)
F2,47 = 1.465
—
SST
9.56 (2.16)
p<.001
p = .028
11.81 (10.3)
p = .242
HC
14.81 (3.19)
13.77 (3.78)
WAIS-III, L-Number
CRT
6.94 (2.19)
F2,47 = 18.08
F1,34 = 1.607
8.76 (2.41)
F2,47 = .812
—
SST
7.88 (2.03)
p<.001
11.69 (11.2)
p = .431
HC
11.88 (3.12)
p = .214
13.31 (2.65)
WAIS-III, Symbol Coding
CRT
50.76 (22.04)
F2,47 = 30.13
F1,34 = .111
60.47 (17.14)
F2,47 = .761
—
SST
52.94 (14.30)
p<.001
p = .741
59.56 (13.67)
p = .474
HC
91.25 (11.51)
93.23 (14.32)
WMS-III, Logical Memory
CRT
25.29 (9.48)
F2,47 = 19.97
F1,34 = 6.568
34.29 (11.11)
F2,47 = 4.006
F1,34 = 5.174
SST
24.75 (11.6)
p<.001
p = .015
37.51 (10.82)
p = .025
p = .030
HC
47.08 (9.05)
49.77 (7.81)
WMS-III, Faces
CRT
31.82 (5.26)
F2,47 = 10.35
F1,34 = 13.28
36.29 (4,96)
F2,47 = 3.065
F1,34 = 4.9
SST
32.75 (11.6)
p<.002
p = .001
37.81 (5.63)
p = .047
p = .034
HC
40.46 (4.77)
43.31 (3.37)
RAVLT
CRT
41.88 (9.84)
F2,47 = 11.15
F1,34 = 1.115
45.41 (11.6)
F2,47 = 1.925
—
SST
45.31 (8.75)
p<.001
p = .299
43.31 (9.98)
p = .158
HC
56.44 (8.93)
59.31 (8.07)
TMT-A
CRT
42.53 (31.96)
F2,47 = 5.197
F1,34 = .932
33.82 (11.8)
F2,47 = 2.033
—
SST
42.09 (13.48)
p<.001
p = .342
36.38 (15.2)
p = .143
HC
24.38 (13.55)
21.23 (6.04)
TMT-B
CRT
128.41 (52.4)
F2,47 = 12.89
F1,34 = 1.801
87.71 (43.66)
F2,47 = 4.698
F1,34 = 1.584
SST
123.44 (54.4)
p<.001
p = .189
81.19 (28.81)
p = .014
p = .218
HC
50.63 (15.36)
57.61 (16.73)
TOL, Total Score
CRT
45.53 (16.24)
F2,47 = 19.19
F1,34 = 3.429
28.29 (13.74)
F2,47 = 19.193
F1,34 = 28.91
SST
43.41 (18.54)
p<.001
p = .074
33.56 (17.28)
p<.001
p<.001
HC
12.62 (12.02)
11.77 (14.18)
WCST, No. of Categories
CRT
2.82 (2.24)
F2,47 = 51.01
F1,34 = 3.782
5.06 (1.88)
F2,47 = 19.193
F1,34 = 15.37
SST
2.25 (1.9)
p<.001
p = .061
4.25 (2.14)
p<.001
p<.001
HC
9.19 (1.76)
8.77 (1.36)
WCST, Perseverations
CRT
24.21 (12.79)
F2,47 = 36.29
F1,34 = .843
24.41 (8.21)
F2,47 = 10.697
F1,34 = 10.68
SST
27.51 (10.3)
p<.001
p = .413
27.13 (8.81)
p<.002
p<.003
HC
1.33 (1.76)
1.56 (1.93)
CRT, cognitive remediation therapy; HC, healthy control subjects; RAVLT, Rey Auditory Verbal Learning Test; SST, social skills training; WAIS-III, Wechsler Adult Intelligence Scale—Third Edition; WMS-III, Wechsler Memory Scale—Third Edition Alternative versions (list A and B) were used in pre- and postassessments; TMT-A, Trail Making Test, part A; TMT-B, Trail Making Test, part B; TOL, Tower of London; WCST, Wisconsin Card Sorting Test.
Figure 2Endpoint-baseline differences through different cognitive domains for each group. The increment was calculated subtracting the baseline scores from the endpoint scores of each cognitive domain. Cognitive domain scores were obtained by calculating the mean of the standardized t scores (mean: 50; SD: 10) from the respective subtests of each cognitive domain as follows: Digit Span, Letter-Number Sequencing and Arithmetic (Wechsler Adult Intelligence Scale-Third Revision). Psychomotor speed (PS): Digit Symbol-Coding (Wechsler Adult Intelligence Scale-Third Revision) and Trail Making Test-A. Verbal memory (VM): Rey Auditory Verbal Learning Test and Logical Memory I and II (Wechsler Memory Scale-Third Edition). Nonverbal memory (NVM): Visual Reproduction I and II and Faces I and II (Wechsler Memory Scale-Third Edition). Executive function (EF): Wisconsin Card Sorting Test Categories and perseverations, Trail Making Test-B, Tower of London. CRT, cognitive remediation therapy; HC, healthy control subjects; SST, social skills training; WM, working memory.
Between-Group Comparison. At baseline, whole-group analysis through IC MELODIC produced 22 ICs: 20 were shared with all groups, and 2 evidenced significant differences between groups. Cluster size, Montreal Neurological Institute coordinates, and Z values of the areas with significant differences between groups are described in Table 3. A detailed description of selection the whole independent component analysis results and the selection criteria used to identify significant independent components are provided in Supplement 1. Specifically, the CRT group showed increased independent component (IC)1 task positive component when compared with HC (z = 1.69; p<.04). IC1-related to the central executive network (CEN), showed increased activations in the bilateral middle and inferior frontal gyrus, left medial frontal gyrus, bilateral inferior and superior parietal lobule, bilateral precuneus, and bilateral middle occipital gyrus, compared with HC participants. IC1 time course fitted the time series task (F = 425.14, p < .000001) and was significant for all the contrasts, 2-back>0-back (z = 10.03, p < .000001), 2-back > rest (z = 21.37, p < .000001), and 0-back>rest (z = 16.46, p<.000001). Compared with SST, the CRT group also exhibited increased activation in the IC8 at rest related to the DMN (z = 1.76; p<.03) in the bilateral anterior cingulate, left middle temporal gyrus, bilateral cingulate gyrus, bilateral precuneus, bilateral inferior occipital gyrus, and bilateral lingual gyrus and its time course fitted the time series task (F = 203.47, p<.000001), being significant for both contrasts; rest>0-back (z = 17.40, p<.000001) and rest>2-back (z = 13.15, p<.000001).
Table 3Peak Activation/Deactivation of Independent Component Analysis Cross-Sectional and Longitudinal Results
Anatomic Location
Cluster Size
MNI Coordinates
Z
x, y, z
Baseline: Cross-sectional Data
IC1 CEN, CRT>HC
Task-related activation
L superior parietal lobe
147,597
116, 60, 124
6.37
R middle frontal gyrus
37,666
48, 164, 96
5.82
Medial frontal gyrus (BA 6)
11,461
92, 144, 116
5.29
Task-related deactivation
R parietal lobe
28,226
68, 92, 144
−4.27
Anterior cingulate (BA 32)
15,101
84, 172, 60
−3.89
L inferior occipital gyrus
5243
120, 28, 68
−4.34
R inferior occipital gyrus
3064
60, 32, 68
−4.15
IC8 DMN, CRT>SST
Rest-related activation
89,926
80, 60, 100
5.77
R precuneus
9739
120, 28, 72
5.34
L inferior occipital gyrus (BA 18)
9208
64, 28, 68
5.02
R inferior occipital gyrus (BA 18)
Rest-related deactivation
L inferior parietal lobe
39,774
136, 88, 128
−5.03
R inferior parietal lobe
3597
44, 84, 124
−3.53
R middle frontal gyrus
3408
60, 128, 132
−3.46
Effect Time×Treatment: Longitudinal Data
IC1 CEN, CRT group
Baseline>Posttreatment
Task-related activation
L superior parietal lobule
70,128
116, 64, 124
5.4
L middle frontal gyrus
11,764
124, 176, 88
3.51
R middle frontal gyrus
9601
48, 168, 92
4.79
Task-related deactivation
L posterior cingulate
7139
90, 69, 88
−3.33
Medial frontal gyrus
7038
88, 164, 52
−3.53
L inferior occipital gyrus
1725
120, 28, 68
−3.13
IC2 DMN, CRT group
Baseline>Posttreatment
Rest-related activation
L precuneus
62,548
92, 70, 108
6.04
L middle temporal gyrus
53,598
136, 56, 100
5.58
Rest-related deactivation
L lingual gyrus
63,885
93, 38, 72
−7.7
L precentral gyrus
51,364
128, 108, 132
−7.79
R superior parietal lobe
18,559
68, 60, 132
−6.22
R middle frontal gyrus
16,035
60, 124, 120
−4.91
BA, Brodmann area; CEN, central executive network; CRT, cognitive remediation therapy; DMN, default mode network; HC, healthy control subjects; IC, independent component; L, left; MNI, Montreal Neurological Institute; R, right; SST, social skills training; TR, repetition time.
With regard to longitudinal results, after treatment, the CRT group showed decreased activation in the IC1 CEN in regions that had shown overactivation at baseline (z = 2.89; p<.001): the bilateral middle and inferior frontal gyrus, anterior cingulate, bilateral inferior and superior parietal lobe, and bilateral precuneus. IC1 time course fitted the time series task (F = 335.07, p = .000001) and was significant for all the contrasts, 2-back>0-back (z = 10.03, p<.000001), 2-back>rest (z = 21.37, p<.000001) and 0-back>rest (z = 16.46, p<.000001). Compared with baseline, the CRT group also exhibited decreased DMN activation after treatment (z = 2.15, p<.015) in the anterior cingulate, cingulate gyrus, bilateral precuneus and cuneus, left middle temporal gyrus, and left supramarginal gyrus (Table 3), and its time course fitted the time series task (F = 405.86, p = .000001), being significant for all the contrasts, rest>0-back (z = 19.09, p<.000001) and rest>2-back (z = 20.71, p<.000001). No significant changes were observed for the HC and SST groups at endpoint. In addition, comparing the pattern of changes between CRT and SST, the CRT group had decreased IC1 CEN activation (z = 1.78, p<.03) indicating a significant between-group interaction (Figure 3). IC1 time course fitted the task; F = 335.07 and p = .000001, and was significant for 2-back>0-back contrast (z = 10.51, p<.00001), 2-back>rest (z = 20.04, p<.00001), and 0-back>rest (z = 14.15, p<.00001).
Figure 3Tensor independent component analysis showing functional changes in the cognitive remediation therapy group after treatment. Changes include a decreased activation of the central executive network (hot colors) and default mode network deactivation (cold colors) during the 2-back task (z = 2.89; p< .001).
Endpoint cross-sectional analysis: neither differences between groups emerged on task-related activation or in the task-related deactivation of the IC1 CEN component at the endpoint. Similarly, IC2 DMN rest-related activation and rest-related deactivation did not show differences between groups on the participants-mode vectors (Supplement 1).
DTI Results. For longitudinal results, by using the TFCE method and p<.05 (FWE corrected), we found that the CRT group showed increase FA values after treatment in the genu and body of the CC, and in the right posterior thalamic radiation (Table 4; Figure S2 in Supplement 1). Conversely, SST showed decreased FA in bilateral superior longitudinal fasciculus, left inferior longitudinal fasciculus (Table 3). No significant changes in FA emerged in the HC group between baseline and after treatment. Moreover, FA changes observed in the CRT group were also significant compared with changes in the SST group (t = 3.27; p<.05). To check that treatment-related changes in FA were not due only to baseline differences, a region of interest analysis between groups was also carried out. A mask based on TBSS longitudinal results was created and mean FA values were extracted from each participants’ FA skeletonized region of interest. Using the TFCE method (p<.05; FWE corrected) a t test between CRT and SST groups with the extracted FA values at baseline was then performed and results showed no significant differences between groups (t = –1.131, df = 28, p = .268).
Table 4Changes in the FA Index in the Patient Groups After Psychological Treatments
Cluster Size
Talairach Coordinates
t
x, y, z
CRT: Increased FA
Genu of CC
619
−5, 23, 15
3.11
Body of CC
122
11 13 24
1.67
Posterior thalamic radiation
10
29, −66, 13
4.11
SST: Decreased FA
Right superior LF
11,173
29, −58, 25
3.29
Left superior LF
6444
−35, −31, 34
1.74
Left inferior LF
125
−38, −48, −5
3.85
All provided results were obtained after being corrected (p< .05).
CC, corpus callosum; CRT, cognitive remediation therapy; FA, fractional anisotropy; LF, longitudinal fasciculus; SST, social skills training.
Structural-Functional Performance Relationship. After treatment, reduction of overactivation in IC1 CEN was found to be associated with improvement in total cognition score (r2 = –.479; p<.03). Moreover, increment in FA was also associated with total cognition score (r2=.442; p<.04), being the executive function domain (r2=.359; p<.029) significantly correlated, but not to the level of the multiple comparisons correction. The association between anatomic and functional changes observed in the CRT group, a posterior analysis was investigated by introducing the IC1 CEN longitudinal subject mode as a covariate in the FA longitudinal analysis. The result, FWE corrected (p<.05), using TFCE showed a negative relationship between increased FA in the CC and decreased CEN overactivation after treatment, suggesting that the white matter integrity increment was associated with normalization of the functional activation pattern during the task.
Discussion
Patients exposed to the treatment showed a reduction in overactivation of the CEN during task-related responses and also in the deactivation of its anticorrelated DMN suggesting an improvement in the efficiency of both networks. On the other hand, an increase in white matter integrity in the genu of the corpus callosum was found in the CRT group after treatment. Functional and structural changes were correlated in the CRT group participants.
Previous studies have already shown a reduction in initial hypofrontality after CRT in patients with schizophrenia following drill-and-practice approaches (
). Additionally, enhancement of activation in prefrontal regions subserving working memory (WM) has also been demonstrated with strategy-based programs that help patients to develop compensatory strategies for learning, remembering, and processing information actively (
Neuroprotective effects of cognitive enhancement therapy against gray matter loss in early schizophrenia: Results from a 2-year randomized controlled trial.
). It needs to be stressed that structural deficits can produce functional deficits at distant regions forming part of the same circuitry.
Conceivably, working memory abnormalities cannot be understood exclusively in terms of the hypofunctioning or hyperfunctioning of the prefrontal areas (
). Inefficiency of more complex neural networks has been postulated for a more descriptive and clarifying framework in which WM abnormalities in schizophrenia are the consequence of an inefficient neural strategy in different areas mediated by the dorsolateral prefrontal cortex (
). Our results can be interpreted in this framework because an initial overactive CEN became more similar to HC participants as a result of the cognitive treatment. On the other hand, normalization of DMN deactivation after treatment might be interpreted in the same context. The DMN is not only important as a resting state; it also seems to have an important role in the performance of cognitive tasks (
). Our study is the first to show that the anomalies in DMN in schizophrenia patients can be ameliorated to some extent with CRT. Although the therapeutic effects on DMN functioning have hardly been studied, at least one previous work showed that treatment with olanzapine was associated with the modulation of DMN connectivity in schizophrenia but not with other prefrontal networks (
) and shed new light on their potential role in cognitive improvement following CRT in schizophrenia. It is conceivable that CRT could have had played a role in cases in which patients who initially were more self-referential (
) provided some early evidence on the neurotrophin response induced by neuroplasticity-based CRT. Schizophrenia participants who followed the computerized cognitive training showed a significant increase in serum brain-derived neurotrophic factor compared with matched control participants. Furthermore, the impact of CRT on structural connectivity over working memory–related networks has never been tested before in patients with schizophrenia, although changes in white matter connectivity have been already described in healthy people after cognitive training. Takeuchi et al. (
) demonstrated that working memory training was associated with increases in white matter structural integrity adjacent to the anterior part of the body of the corpus callosum in healthy people. The anterior part of the corpus callosum connects the dorsolateral prefrontal cortex between both hemispheres, and the whole network is a key element of the working memory system (
). Those improvements will lead to an increase of interhemispheric information transfer between the bilateral prefrontal cortices. Surprisingly, our findings showed similar results on the same brain structure with schizophrenia patients, suggesting a common underlying mechanism for the induced neural plasticity. It has been speculated that cognitive training could be facilitating specific neural activity that would eventually cause increased myelination in fiber tracts (
In summary, our results show that there are detectable effects on functional and structural connectivity after CRT in patients with schizophrenia. Despite the beneficial effects demonstrated in this study, these findings need to be interpreted in the context of a number of important limitations. In all likelihood, the current pretest/posttest design is, unfortunately, too weak to fully characterize the dynamics of change over time. Moreover, the version of the cognitive task we used in the scanning (0-back vs. 2-back comparison) has not allowed us to characterize properly the clinical improvement due to CRT. Other obvious problem is the role of pharmacologic treatment in cognitive improvement. An important challenge for clinical research will be to investigate combinations of cognitive training with specific pharmacologic treatments (
), it is unclear from this study what the long-term consequences of training in brain functions might be.
This study was supported by Fondo de Investigación sanitaria (FIS) Grant No. PI 07/0258 (to RP) and Rio Hortega Institut d'Investigació Biomèdica Agust Pi i Sunyer (IDIBAPS) (NP) from the Instituto de Salud Carlos III.
The authors report no biomedical financial interests or potential conflicts of interest.
Clinicaltrials.gov: Effects of Cognitive Remediation Therapy on Schizophrenia Patients Through Functional Magnetic Resonance Imaging; http://clinicaltrials.gov/ct2/show/NCT01318850; NCT01318850.
What are we changing with neurocognitive rehabilitation? Illustrations from two single cases of changes in neuropsychological performance and brain systems as measured by SPECT.
Preliminary evidence of improved verbal working memory performance and normalization of task-related frontal lobe activation in schizophrenia following cognitive exercises.
Neuroprotective effects of cognitive enhancement therapy against gray matter loss in early schizophrenia: Results from a 2-year randomized controlled trial.
First MB, Spitzer RL, Gibbon M, Williams JB (1997): Structured Clinical Interview for DSM-IV Axis I Disorders—Clinician Version (SCID-CV) [Spanish Edition: Barcelona: Masson SA, 2001]. Washington, DC: American Psychiatric Press
Six years ago, at the first Cognitive Neuroscience Treatment Research to Improve Cognition in Schizophrenia (CNTRICS) meeting, a neuroscientist questioned whether attention dysfunction was malleable in schizophrenia, despite a recent report that patients were 5 times more likely to work when cognitive remediation was combined with supported employment (1). The idea that impaired neural systems could demonstrate learning-induced plasticity was not part of the biological research lexicon at that point in time.
Erratum to: “Brain Effects of Cognitive Remediation Therapy in Schizophrenia: A Structural and Functional Neuroimaging Study” by Penadés et al. which appeared in Biological Psychiatry (2013;73:1015–1023).
00062-0&id=gr1.jpg){kind=link}
00062-0&id=gr2.jpg){kind=link}
00062-0&id=gr3.jpg){kind=link}