The cortical and subcortical substrates of quality of life through substrates of self-awareness and executive functions, in chronic moderate-to-severe TBI

ABSTRACT

Background

Moderate-to-severe traumatic brain injury (TBI) follows a chronic neuro-psychological sequelae, interfering with quality of life (QOL).

Objective

To investigate the chronic effects of moderate-to-severe TBI as expressed by greater atrophy in specific regions-of-interest relating to executive functions (EF) and self-awareness (SA); and whether this atrophy reflects on EF, SA deficits and QOL.

Methods

Thirty-one males with chronic moderate-to-severe TBI, aged 18–51, were compared to 24 non-injured males (age range = 21–49), matched on age and education. EF was assessed through a composite score. SA and QOL were assessed using generic and TBI-specific measures. Online masks were applied on magnetic resonance images to extract EF and SA – related regions-of-interest.

Results

Findings revealed that participants with TBI presented with less volume in fronto-temporal cortical and subcortical regions, than controls. An interrelation between EF and SA – related regions was revealed. Participants with TBI scored lower on neuropsychosocial measures, than controls. Differences in EF and SA were reflected on the related regions-of-interest. Satisfaction with QOL was predicted by these regions-of-interest.

Conclusion

Chronic TBI effects on brain volume extend on EF, SA, and QOL; highlighting the role of SA between EF and QOL, and the need for personalized interventions in improving recovery outcome.

KEYWORDS:

• Imaging
• neuropathology
• outcome

Introduction

Traumatic brain injury (TBI) is a major cause of hospitalization, death, and chronic disability, globally (1,2). TBI is accompanied by long-term and progressive disabilities, such as significant neuropsychological impairment (2–6), including executive dysfunction (7,8), and deficits in self-awareness (SA; 2,8,9). Neuropsychological impairment has been associated with chronic and progressive brain volume loss (5,7,10,11), as TBI can result in a pathophysiologic sequelae. This pathophysiologic sequelae is analogous to the location and severity of the damage, diffuse effects, and secondary mechanisms of injury, leading to focal and diffuse brain injury. Contusions can directly disrupt function in both cortical and sub-cortical regions, with areas such as frontal and anterior temporal, and the hippocampus, respectively, being more vulnerable to the trauma due to their position within the skull (12–14).

More specifically, greater injury severity has been found to lead to an increase in pathophysiology in both gray and white matter volume within the frontal lobes and related brain areas and circuits, which further relates with impairment in executive functions (EF), and SA. EF and SA deficits have been shown to interfere with one’s ability to function adaptively (15,16) and their perception of their quality of life (QOL; 8,17). Executive dysfunction and impaired SA have been linked to greater gray and white matter pathology in numerous frontal, temporal, parietal, and occipital cortical and sub-cortical regions, including the cingulate cortex, the medial prefrontal cortex (MPFC), the dorsolateral prefrontal cortices (DLPFC), the superior frontal gyri, the ventrolateral prefrontal cortex, the hippocampus, the thalamus, the insula and the caudate (18–22). Despite clear evidence supporting the link between EF and SA with the aforementioned areas, as well as the associations between EF and SA deficits with QOL, no study has investigated whether the impact of TBI on QOL is reflected in EF and SA – related brain regions. Therefore, the aim of this study was to explore the link between the pathophysiologic effects of TBI in brain areas relating to EF and SA with the individuals’ QOL.

Brain atrophy, executive dysfunction, and impaired self-awareness

Literature has come to a consensus concerning the implication of the frontal lobes in EF, SA, and thus one’s daily participation and QOL. However, Prigatano (15) also highlighted the involvement of other brain areas and circuits in effectively employing EF-related behaviors such as planning, coordinating and monitoring behavior. These neurocircuits have been described in detail by Constantinidou et al. (23), arguing their implication in the relationship of the frontal lobes and EF, through transferring information generated by the cortex and moving this toward subcortical regions, such as the globus pallidus, and the thalamus, and back to the cortex. The first two circuits involve mainly motor functions; whereas the rest are implicated in cognitive and behavioral functions: the dorsolateral prefrontal circuit (DLPFc), the orbitofrontal circuit (OFc), and the anterior cingulate circuit (ACc). Similar behavioral patterns may be produced as the result of damage to these circuits, as the basal ganglia have been linked to the cortex through these circuits (23).

Specifically, damage to the DLPFc is accompanied by greater deficits in organizational skills, greater difficulty in shifting attention, and environmental dependency. Such symptoms have been linked to executive dysfunction, and thus the DLPFc has been described as the circuit relating mostly to EF. Personality changes, including irritability, lack of empathy, and inappropriate social behavior, following a TBI have been attributed to impairment within the OFc. Despite such neural damage to the OFc, individuals may still perform adequately on EF tasks. Finally, damage to the ACc has been reported to result in lack of motivation and apathy, including akinetic mutism, poor response inhibition, minimal creative thinking, and difficulty in producing spontaneous speech (see 23).

In a study investigating the relationships between regional variation in gray matter volume and cognitive impairment in individuals with mild to severe TBI, evidence supported the association between the two, whereby the participants with most severe injuries displayed the most significant impairment (22). Specifically, patients with lower scores on executive control performance indicated lower cortical volume in temporal, parietal, and occipital regions. Therefore, it may be argued that behavioral deficits relating to damage to the EF system may result from damage to the frontal and prefrontal regions, as well as to the cortical and subcortical regions connected to this system, including temporal, parietal, and occipital areas (23). As an effect, research has focused on further exploring the underlying networks guiding such behavior.

Similar brain areas and connections have been reported to relate to impaired SA (15,20). In a review article, the authors posit the implication of two circuits in a number of processes depicting SA: 1) the DLPFc “(responsible for self-regulation, self-monitoring, and other EF), which includes links to the basal ganglia, thalamus and prefrontal cortex,” and 2) the OFc “(responsible for empathic and socially appropriate responses) that includes links to the basal ganglia, thalamus, and orbitofrontal cortex” (see review 24). Evidence has shown the implication of the anterior cingulate cortex in procedural learning and behavior modification through transferring reinforcing stimuli to diffuse areas of cortical and subcortical regions. In addition, O’Connell et al. (25,26) have reported the implication of the anterior cingulate cortex in error diagnosis and detection, and conflict processes. Such processes are enabled when one is called to respond correctly at the presence of competitive stimuli. The aforementioned processes signal self-monitoring and regulation abilities, which directly relate to emergent awareness. Taylor et al. (20) support this evidence and highlight the involvement of a number of neural regions in error awareness, including the dorsal and rostral anterior cingulate cortex, the posterior and anterior MPFC, and the prefrontal cortex. These areas allow for the development of adaptive behaviors, that in turn further strengthen global awareness (19).

These findings clearly support the association between TBI neuropathology with impaired EF and SA. It is also well known that deficits in EF and SA relate to health-related QOL outcome, with individuals with moderate-to-severe TBI over-reporting their QOL levels (see 8), mainly regarding their cognitive abilities (17). However, no evidence exists to describe the relationship between neuropathology in these EF and SA – related areas and health-related QOL. Therefore, this study has sought to investigate whether greater atrophy in these areas extends to the over-reporting of one’s QOL levels. Highlighting the chronic and persistent course of both global and regional brain volume loss, and specifically in EF and SA – related regions of interest, and how these may further reflect in one’s QOL and health-related QOL (for a definition and differentiation of these constructs see 8); will help with the development of related biopsychosocial models and the designing of more comprehensive rehabilitation programs by focusing on the impact of impairments on daily participation. This would further improve patient recovery.

Specifically, it was expected that participants with chronic moderate-to-severe TBI with greater atrophy in EF and SA -related regions of interest (ROIs), including the MPFC, the cingulate cortex, the thalamus, the caudate, the temporal lobe, and the cerebellum would also present with greater EF and SA deficits. Finally, given the associations between EF and SA with QOL, it was hypothesized that QOL and health-related QOL would negatively correlate with EF and SA -related ROIs in this participant group.

Materials and methods

Participants

This study was part of a larger project exploring the effects of chronic moderate-to-severe TBI. Participants with TBI were recruited from the Nicosia General Hospital Intensive Care Unit, and the Melathron Agoniston EOKA databases. All study procedures have been approved by the Cyprus National Bioethics Committee, and all participants provided written informed consent. All participants underwent a comprehensive pen-and-paper neuropsychological and psychosocial assessment, along with MRIs.

The group with TBI consisted of 31 Greek-speaking males with a primary diagnosis of moderate-to-severe closed-head injury (see Supplementary Material for inclusion/exclusion criteria). Participants with TBI had an age range of 18–51 years old (M = 31.48, SD = 8.54), a mean educational level of 12.55 years (SD = 2.97; range = 6–19), and were assessed at a mean time since injury of 5.48 years (SD = 5.73; range = 1–19). TBI causes were consistent with literature (1), with the primary cause being motor vehicle collisions, accounting for 83.9% of the sample. Most participants with TBI had received acute inpatient rehabilitation; but three had received fragmented individualized outpatient treatment, with no participant having received post-acute rehabilitation services (for a full description of each participant, see Table 1).

Table 1. Demographic information of the participants with TBI.

Participant No.	Age (Years)	Education (Years)	Cause of Injury	TSI (Years)	PTA (days)	LoHS (days)
1	21	11	MVC	2	30	61
2	24	12	MCC	2	N/A	88
3	31	8	MCC	4	30	30
4	51	6	PVC	9	183	30
5	44	8	Fall or Work-related Injury	6	21	45
6	26	13	MVC	8	3	N/A
7	37	12	PVC	2	2	18
8	23	12	MCC	3	20	60
9	30	9	MCC	4	80	122
10	26	14	MCC	5	183	10
11	38	12	Object Falling	1	25	53
12	44	12	Fall or Work-related Injury	1	61	61
13	34	12	MVC	2	30	30
14	45	15	Assault	17	N/A	N/A
15	22	15	PVC	4	7	30
16	21	12	MCC	2	28	30
17	18	11	MCC	4	244	95
18	34	12	MVC	1	N/A	N/A
19	43	19	PVC	19	42	42
20	29	11	MCC	4	335	N/A
21	25	13	MCC	3	25	60
22	32	18	MCC	18	22	30
23	23	12	MVC	2	30	30
24	28	16	MVC	1	6	30
25	20	12	MVC	1	7	56
26	31	8	MVC	1	61	122
27	33	16	MVC	1	61	61
28	40	14	Assault	14	3	30
29	29	16	MCC	1	61	61
30	36	12	MVC	17	15	180
31	38	16	MCC	11	335	N/A

MVC, Motor Vehicle Crash; MCC, Motor Cycle Crash; PVC, Pedestrian – Vehicle Collision; TSI, time since injury; PTA, Posttraumatic Amnesia; LoHS, Length of Hospitalization Stay.

Participants with TBI were matched to 24 healthy Greek-speaking individuals on gender, and age, and education with a variance of ± two years (age, M = 31.92 years; SD = 8.18; range = 21–49; education, M = 13.63 years; SD = 2.48; range = 8–17). The neurotypical group consisted of volunteers recruited from the greater areas of Cyprus. Healthy controls did not report any history of TBI or any other neurological condition, documented psychological disorder, substance abuse, or learning disability.

The two groups were very similar in terms of age and education (age, t (53) = −0.19, p = .850; education, t (53) = −1.43, p = .158). Therefore, any significant differences in subsequent analyses cannot be attributed to sample differences.

Imaging protocol

For image acquisition a 3.0 Tesla scanner (Achieva, Philips Medical Systems, Best, The Netherlands) was used. The built-in quadrature RF body coil and a phased array 8-channel head coil was used for proton excitation and signal detection, respectively. An isotropic, three-dimensional (3D), T1-weighted rapid acquisition gradient-echo sequence (fast field echo; repetition time = 25 ms; echo time = 1.85 ms; flip angle = 30o) allowed for acquiring whole brain, transverse MR images with an acquisition/reconstruction voxel of 1.0 × 1.0 × 1.0 mm (data interpolation was not implemented in any direction to improve resolution and reduce partial volume effects). The scanning session included other standard pulse sequences to exclude significant brain pathology of a different etiology.

Executive functions

For executive functions (EF), standard score transformations of commonly used speeded measures of EF were conducted, mainly assessing attentional control, verbal fluency and set-shifting abilities, to create a unified construct reflecting different elements of EF. Specifically, (i) individual test scores (see Table 2) were reversed, where needed, with lower scores indicating worst performance. (ii) These were then transformed into standard scores (z-scores), based on the mean of the neurotypical participants. Finally, (iii) each test’s z-scores were averaged to obtain a single score measuring EF. The validity of this composite score for this study is illustrated in Table 2, indicating high correlations between each task score and the EF composite score.

Table 2. Correlations between individual z-scores of EF tasks and the EF composite score.

	1	2	3	4	5	6	7
1. Rey Copy	-
2. TMT A	0.69**	-
3. TMT B	0.66**	0.68*	-
4. SDMT	0.66**	0.70**	0.65**	-
5. COWAT Animals	0.48**	0.39**	0.33**	0.52**	-
6. COWAT Letter F	0.36**	0.35**	0.34**	0.43**	0.31*	-
7. Executive Functions Composite	0.75**	0.83**	0.79**	0.87**	0.67**	0.65**	-

p < 0.01, one-tailed

Self-awareness

SA was assessed via the Greek version of the DEX–R, which is a 37-item inventory with three derived factors: (i) Motivation and Attention, (ii) Flexibility, Fluency, and Working Memory and (iii) Social and Self -Regulation. As proposed by Prigatano (27; e.g. 17), SA was computed by measuring the discrepancy between the participant score and the informant score, i.e. DEX-R-Discrepancy_i = DEX-R-Participant – DEX-R-Informant, for each scale, with negative scores indicating worse SA.

DEX-R has been criticized as a measure of self-awareness, as patient and informant scores do not differ significantly in all studies, and the DEX correlates strongly with depression severity. Therefore, in this study it was used to delineate SA deficits in participants with TBI over neurotypical participants. However, for the participants with TBI, the Self-Regulation Skills Interview (SRSI; 28) was also completed. The SSRI is a five-item semi-structured interview measuring emergent awareness, anticipatory awareness, strategy generation, strategy-use, and strategy effectiveness, that is specific to TBI. The five items are scored on a 10-point scale, with higher scores representing lower levels of SA. Upon receiving author’s permission, the scale was translated in Greek via the forward and backwards method by a professional translator.

Quality of life

Two questionnaires were employed as measures of QOL, in an attempt to follow the ICF (29) concepts to thoroughly describe the health-related QOL phenomenon. Therefore, both a generic and a TBI-specific measure were used: (i) The Greek version of the World Health Organization Quality of Life assessment instrument-BREF (WHOQOL-BREF; 30), a 26-item questionnaire assessing a person’s subjective perception of their life regarding to their goals, concerns, and satisfaction on five factors: physical and psychological health, social relationships, environmental aspects and one overall well-being, thus covering a sufficient number of concepts of the ICF. Items were scored on a 5-point response scale (1 – “Very poor” and 5 – “Very good”), with lower scores indicating lower satisfaction with QOL.

(ii) The Quality of Life after Brain Injury (QOLIBRI; 31), a 37–item measure yielding six domains pertaining to health-related QOL: cognition, self, daily life and autonomy, social relationships, emotions and physical problems. This measure also covers multiple aspects of the ICF that are TBI-specific. All items are rated on a 5-point Likert scale (1 – “None” and 5 – “Very much”), with lower scores indicating lower health-related QOL satisfaction.

Statistical analysis

Prior to any analyses, pre-processing of the MR images was conducted using SPM12. Pre-processing steps included segmentation of the MR images into GM and WM, and CSF, followed by a Diffeomorphic Anatomical Registration through Exponentiated Lie Algebra (DARTEL) for inter-subject registration of the GM, WM, and CSF images. Local GM, WM, and CSF volumes were conserved by modulating the image intensity of each voxel by the Jacobian determinants of the deformation fields computed by DARTEL. The registered images were, then, smoothed with a Gaussian kernel (Full Width at Half Maximum = 8 mm) and were further transformed to Montreal Neurological Institute (MNI) stereotactic space using affine and nonlinear spatial normalization implemented in SPM12 for statistical comparisons.

Volumetry

Volumetry was used to detect group differences in overall GM, WM, and CSF volume, using IBASPM to calculate individual brain volume. These indexes allowed for the quantification of tissue. Indexes were entered into SPSS and independent samples t-tests were conducted to compare the two groups.

Voxel-based-morphometry

Voxel-based-morphometry analyses were conducted to investigate whether significant volume reduction in whole-brain regions was evident between the two groups. These hypotheses were tested through conducting independent samples t-test models in SPM12, with age, education, and overall brain volume entered as covariates of no interest.

Regions-of-interest

Further analyses were performed to investigate the associations between the volume in EF and SA – related brain regions and neuropsychological performance, and psychosocial measures. Specifically, regression analyses were performed to investigate the predictive validity of brain volume in regions-of-interest (ROIs) in executive dysfunction, impaired SA and QOL. Therefore, masks of brain regions relating specifically to EF, and SA were downloaded from the database of Neurosynth.org. These maps were then entered into MRICRON, and individual masks of each ROI were hand-drawn. Each mask was then used to extract the volume from each ROI, using MATLAB. All data were entered into SPSS. Prior to examining the main hypotheses, group differences in volumetry, and EF, SA and QOL measures were tested using t-tests, and in ROIs using MANCOVA, with age, education and overall brain volume entered as covariates of no interest. Finally, correlation and stepwise regression analyses were performed to investigate the predictive value of the ROIs on EF, SA, and QOL. Corrections for multiple comparisons were applied by lowering the α-level to 0.01.

Results

Volumetry

Independent-samples t-tests revealed that the two groups significantly differed in mean GM volume, t (53) = −3.20, p = .002, Cohen’s d = 0.85, with participants with TBI presenting with less GM volume (M = 635.62 cm³, SD = 63.22 cm³) as compared to the control group (M = 701.26 cm³, SD = 88.94 cm³). Similar findings were also shown for mean WM volume with participants with TBI showing significantly less volume (M = 392.51 cm³, SD = 74.56 cm³) than the non-injured individuals (M = 495.72 cm³, SD = 59.98 cm³), t (53) = −5.53, p = .0001, Cohen’s d = 1.53.

As an effect of the reduction in both GM and WM volume, the CSF volume was significantly larger in participants with TBI (M = 367.53 cm³, SD = 76.19 cm³), than the control group (M = 273.93 cm³, SD = 60.16 cm³), t (53) = 4.94, p = .0001, Cohen’s d = 1.36.

Voxel-based morphometry

Whole-brain analysis was further conducted to distinguish between brain areas that significantly differed in GM and WM volume between the two groups. Participants with TBI showed significantly less volume in GM in the left MPFC, the left middle frontal gyrus and the right cerebral WM, as compared to the neurotypical group (see Table 3). No brain regions were significantly larger in volume in participants with TBI, as compared to the neurotypical group.

Table 3. Whole-brain VBM analysis.

Anatomical Region	Tissue Type	Side	MNI Coordinates			Peak-z	p-value
			x	y	z
Medial Frontal Cortex	GM	L	−2	33	−18	7.67	0.002
Medial Frontal Gyrus	GM	L	45	−22	0	7.62	0.001
Cerebral	WM	R	30	57	3	7.36	0.002

p < 0.01, one-tailed

Table 4. Group differences in EF and SA-related brain regions.

	Group with TBI			Control Group			Statistics
Executive Functions	Side	M	SD	M	SD	F	P	η^2	Observed Power
EF-related Areas
Caudate	L & R	0.20	0.04	0.24	0.04	21.36	0.000	0.30	0.99
Cingulate Cortex	R	0.44	0.07	0.51	0.09	13.39	0.001	0.21	0.95
Globus Pallidus	L	0.08	0.01	0.09	0.01	29.48	0.000	0.37	1.00
Insula	L & R	0.45	0.06	0.54	0.07	32.15	0.000	0.39	1.00
MPFC	L & R	0.33	0.04	0.36	0.05	11.56	0.001	0.19	0.92
OFC	L & R	0.35	0.05	0.43	0.06	33.31	0.000	0.40	1.00
Putamen	L & R	0.27	0.03	0.31	0.05	17.02	0.000	0.25	0.98
Temporal Cortex	L & R	0.45	0.06	0.51	0.06	24.68	0.000	0.33	1.00
Temporal Pole	R	0.40	0.08	0.47	0.06	17.32	0.000	0.26	0.98
Thalamus	L	0.40	0.08	0.49	0.10	15.34	0.000	0.23	0.97
SA-related Areas
Basal Ganglia	L & R	0.23	0.03	0.27	0.04	30.61	0.0001	0.38	1.00
Cingulate Cortex	L & R	0.44	0.05	0.53	0.08	29.62	0.0001	0.37	1.00
Corpus Callosum	L & R	0.27	0.04	0.33	0.05	34.00	0.0001	0.40	1.00
Globus Pallidus	R	0.10	0.01	0.11	0.01	15.65	0.0001	0.24	0.97
Hippocampus	L & R	0.51	0.06	0.58	0.05	30.27	0.0001	0.38	1.00
Inferior Frontal Gyrus	L & R	0.48	0.06	0.53	0.06	10.77	0.002	0.18	0.90
Insula	L	0.50	0.05	0.57	0.08	22.35	0.0001	0.31	1.00
MPFC	L & R	0.35	0.04	0.43	0.07	31.74	0.0001	0.39	1.00
OFC	L & R	0.41	0.06	0.52	0.09	35.87	0.0001	0.42	1.00
Putamen	L	0.39	0.04	0.45	0.05	31.86	0.0001	0.39	1.00
Temporal Area	L & R	0.19	0.04	0.23	0.06	15.22	0.0001	0.23	0.97
Temporal Cortex	L & R	0.42	0.06	0.49	0.06	28.15	0.0001	0.36	1.00
Temporal Pole	L & R	0.41	0.07	0.47	0.06	18.06	0.0001	0.27	0.99

p < 0.01; Group with TBI, N = 31; Control group, N = 24.

Regions–of–interest

Neural systems involved with the concepts of EF and “Self” have been detected in healthy adults and are presented in the Neurosynth Database. A meta-analyses brain map of 97 studies for EF and one of 903 studies for the “Self” concept were extracted from the NeuroSynth database, revealing neural substrates involving these two concepts separately (http://www.neurosynth.org; 32). The map relating to EF identifies clusters mainly in the bilateral frontal and temporal cortical and subcortical regions. Similar areas, including the parietal cortex, seem to be involved with the term “self.” Therefore, specific ROIs were investigated for group differences in EF and Self -related structures, using the meta-analyses masks extracted from Neurosynth.org.

Executive functions

Multivariate ANCOVA was conducted with age, educational level, and global volume entered as covariates. Also, to account for multiple statistical tests, the α level was reduced to 0.01. A group effect was found on the EF brain regions, F (20, 31) = 3.05, p = .003, η² = 0.66, Observed Power = 0.99, with participants with TBI showing significantly less volume in a great number of cortical and sub-cortical areas involved in EF, as compared to the non-injured controls, including the left globus pallidus, the orbitofrontal cortex (OFC), the putamen, the temporal cortex, the right temporal pole, the left thalamus, the insula, the caudate, the right cingulate cortex, and the medial prefrontal cortex (MPFC; Table 4 and Figure 1).

Figure 1. Differences in EF-related brain regions.

Self-Awareness

A similar analysis was performed to investigate differences in regions relating to the concept of “Self,” between the two groups. Again, to account for multiple statistical tests, the α level was reduced to 0.01. When controlling for age, education, and overall brain volume, differences in areas relating to the “Self” were detected between the two groups, F (25, 26) = 2.33, p = .008, η² = 0.69, Observed Power = 0.95. Specifically, participants with TBI displayed significantly less brain volume in numerous regions, such as the OFC, the cingulate cortex, the temporal cortex and pole, the inferior frontal gyrus, the MPFC, and the hippocampus, bilaterally, as well as the left insula and putamen, the pons, and the right globus pallidus, compared to the control group (Table 3 or Figure 2 for graphic display).

Figure 2. Differences in SA-related brain regions.

Neuropsychological and psychosocial measures

Independent-samples t-test revealed a significant difference on the EF construct, t (42.48) = −5.83, p = .0001, Cohen’s d = 1.55, with participants with TBI underperforming (M = 0.55, SD = 0.33), as compared to the neurotypical group (M = −0.37, SD = 0.77). Participants with TBI presented with greater deficits on social and self -regulation awareness (M = −12.36, SD = 14.49), t (35.73) = −4.42, p = .0001, Cohen’s d = 1.14; motivation and attention awareness (M = −4.13, SD = 6.63), t (33.58) = −3.91, p = .0001, Cohen’s d = 1.00; and overall awareness (M = −19.81, SD = 29.41), t (33.51) = −3.77, p = .001, Cohen’s d = 0.96; as compared to the controls (social and self – regulation, M = −0.29, SD = 3.99; motivation and attention, M = 0.67, SD = 1.44; overall awareness, M = 0.67, SD = 6.30). Finally, independent-samples t-test revealed group differences for the environmental, t (53) = 2.15, p = .031, Cohen’s d = 0.59, and physical, t (48.49) = −2.12, p = .039, Cohen’s d = 0.56, aspects of QOL. Participants with TBI reported greater satisfaction with the environmental aspects of their lives (M = 31.71, SD = 4.03) and less satisfaction with the physical ones (M = 27.36, SD = 4.86), than the neurotypical group (environmental, M = 29.00, SD = 5.06; physical, M = 29.54, SD = 2.69).

ROIs

One-tailed Pearson’s correlations were performed to explore whether the EF-related regions correlated with the SA-related regions. Corrections for multiple comparisons were applied by lowering the α-level to 0.01. As expected, EF and SA – related brain areas showed significant associations (Table 5).

Table 5. Correlations between EF-related regions & SA-related regions.

	SA
											Temporal
	Basal Ganglia	CC	Corpus Callosum	Globus Pallidus	Hippocampus	IFG	Insula	MPFC	OFC	Putamen	Cortex	Pole	Area
EF
Caudate	0.60*	0.65*	0.68*	0.66*	0.66*	0.60*	0.54*	0.65*	0.61*	0.56*	-	0.56*	-
CC	0.80*	0.75*	0.76*	0.66*	0.61*	0.75*	0.65*	0.82*	0.81*	0.76*	0.55*	0.76*	0.48*
Globus Pallidus	0.70*	0.65*	0.76*	0.80*	0.59*	0.69*	0.72*	0.69*	0.63*	0.55*	-	0.59*	-
Insula	0.88*	0.76*	0.85*	0.83*	0.62*	0.69*	0.91*	0.83*	0.85*	0.66*	0.58*	0.68*	-
MPFC	0.74*	0.71*	0.71*	0.62*	0.61*	0.58*	0.62*	0.85*	0.76*	0.68*	0.63*	0.74*	0.51*
OFC	0.90*	0.77*	0.80*	0.76*	0.55*	0.64*	0.79*	0.84*	0.93*	0.65*	0.52*	0.67*	0.37*
Putamen	0.71*	0.65*	0.73*	0.79*	0.70*	0.72*	0.68*	0.70*	0.64*	0.57*	0.42*	0.60*	0.42*
Temporal Cortex	0.74*	0.81*	0.73*	0.59*	0.62*	0.52*	0.65*	0.73*	0.69*	0.53*	0.53*	0.69*	-
Temporal Pole	0.72*	0.69*	0.67*	0.58*	0.66*	0.48*	0.60*	0.66*	0.67*	0.48*	-	0.61*	-
Thalamus	0.65*	0.60*	-	0.61*	0.59*	0.49*	0.60*	0.55*	0.58*	-	-	0.59*	0.46*

*p < 0.001, one-tailed; N = 53, EF, Executive Functions; SA, Self-awareness; CC, Cingulate Cortex; IFG, Inferior Frontal Gyrus; MPFC, Medial Prefrontal Cortex; OFC, Orbitofrontal Cortex.

ROIs and neuropsychosocial measures

Given the group differences detected in the neuropsychological and psychosocial measures, that is, EF, SA, and QOL, it was sought to examine whether these differences were reflected on the neural substrates involving these behaviors, for the areas which participants with TBI exhibited lower volume compared to the neurotypical group. For this reason, one-tailed Pearson correlational analyses were conducted to investigate the relationships between the ROIs, EF and SA, and the psychosocial measures. For the variables displaying significant correlations (r ≥ ±0.4; see Table 6 for EF-ROIs correlations, and Table 7 for SA-ROIs correlations), stepwise regression analyses were conducted to examine the predictive validity of the ROIs on EF, SA, and QOL. Corrections were applied to correlations and regressions due to numerous comparisons, by lowering the α-level to 0.01.

Table 6. Correlations between EF-related regions and neuropsychosocial measures.

	EF-Related Areas
	MPFC	Globus Pallidus (L)	OFC	Putamen	Temporal Cortex	Temporal Pole	Thalamus (L)	Caudate	CC (R)
EF Construct	0.49**	0.39*	-	.43**	0.36*	-	0.43**	0.52**	0.56**
DEX-R
F, F, WM	-	-	0.44**	-	-	0.52**	-	0.35*	0.33*
S & S-R	-	-	0.40**	-	0.13	0.42**	-	-	-
M & A	-	0.34*	0.38*	0.32*	0.41**	0.55**	-	-	0.31*
Overall SA	-	-	0.44**	-	-	0.51**	-	-	-
SRSI
Strategic Awareness	−0.52**	−0.37*	−0.39*	−0.54**	−0.39*	−0.50**	−0.39*	−0.53**	−0.52**
Online Awareness	−0.45**	−0.42**	−0.41**	−0.51**	−0.31*	−0.46**	−0.37*	−0.53**	−0.47**
WHOQOL-BREF
Psychological Health	-	−0.33*	−0.37*	-	-	−0.32*	-	-	-
Social Relationships	-	-	−0.36*	-	-	-	-	-	-
Environment	−0.33*	-	−0.34*	−0.38*	-	−0.40*	-	-	−0.41**
Overall QOL	-	-	−0.36*	-	-	-	-	-	-
QOLIBRI
Cognition	-	−0.43**	-	-	-	-	-	−0.33*	-
Self Perception	−0.40**	−0.40**	−0.38*	−0.31*	-	−0.38*	-	−0.32*	−0.37*
DL&A	−0.34*	-	-	-	-	-	-	-	-
Social Relationships	−0.40*	-	−0.45**	-	-	-	-	−0.31*	−0.37*
Emotions	-	-	−0.42**	-	-	-	−0.39*	-	-

**p < 0.01; *p < 0.05; one-tailed; Group with TBI, N = 31; CC, Cingulate Cortex; IFG, Inferior Frontal Gyrus; MPFC, Medial Prefrontal Cortex; OFC, Orbitofrontal Cortex; Flexibility, Fluency & Working Memory,; S & S-R, Social & Self -Regulation; M & A, Motivation & Attention; DL&A, Daily Life & Autonomy.

Table 7. Correlations between SA-related regions and neuropsychosocial measures.

	SA-related regions
					Temporal
	Basal Ganglia	OFC		Insula (L)	Cortex	Pole	Area
DEX-R
F, F & WM	0.35*	0.39*		-	0.44**	0.34*	-
S & S R	-	0.39*		-	0.35*	-	-
M & A	0.39*	0.33*		0.43**	0.48**	0.41**	0.33*
Overall SA	0.35*	0.40**		-	0.43**	0.36*	-
SRSI
Strategic Awareness	−0.33*	−0.35*		−0.37*	−0.50**	−0.48**	−0.42**
Online Awareness	−0.32*	−0.34*		−0.37*	−0.50**	−0.51**	−0.55**
Readiness to Change	-	-		-	-	-	-
WHOQOL-BREF
Psychological Health	−0.32*	−0.31*		−0.35*	-	−0.31*	−0.51**
Social Relationships	-	-		−0.30*	−0.35*	-	-
Environment	−0.41**	-		−0.45**	−0.37*	-	−0.43**
Overall QOL	−0.33*	-		−.33*	-	-	−0.45**
QOLIBRI
Cognition	-	-		−0.34*	-	−0.35*	−0.45**
Self Perception	−0.41**	−0.44**		−0.50**	−0.54**	−0.53**	−0.59**
DL & A	-	−0.33*		-	-	-	−0.32*
Social Relationships	-	−0.44**		-	−0.40*	−0.35*	-
Emotions	-	−0.31*		-	-	-	-
Overall HRQOL	−0.33*	−0.40**		−0.37*	−0.37*	−0.35*	−0.52**
**p < .01; *p < .05; one-tailed; Group with TBI, N = 31; OFC, Orbitofrontal Cortex; Flexibility, Fluency & Working Memory,; S & S-R, Social & Self -Regulation; M & A, Motivation & Attention; DL&A, Daily Life & Autonomy.
	SA-related regions
	IFG	Globus Pallidus (R)	Hippocampus		CC	Corpus Callosum	MPFC
DEX-R
F, F & WM	-	0.30*	0.34*		-	-	0.32*
S & S R	0.33*	-	-		-	-	-
M & A	0.36*	-	0.39*		-	0.32*	-
Overall SA	0.34*	-	0.32*		-	0.31*	-
SRSI
Strategic Awareness	−0.49**	−0.37*	−0.30*		−0.54**	−0.39*	−0.52**
Online Awareness	−0.45**	−0.35*	-		−0.48**	−0.36*	−0.44**
Readiness to Change	-	−0.50**	−0.38*		-	-	-
WHOQOL-BREF
Psychological Health	-	-	−0.34*		−0.36*	−0.33*	-
Social Relationships	−0.33*	-	-		−0.33*	-	-
Environment	−0.13	−0.38*	−0.45**		−0.38*	−0.31*	-
Overall QOL	-	-	−0.31*		−0.36*	-	-
QOLIBRI
Cognition	−0.34*	-	-		-	-	-
Self Perception	−0.43**	−0.33*	−0.50**		−0.53**	−0.46**	−0.44**
DL & A	-	-	-		−0.30*	−0.30*	−0.37*
Social Relationships	−0.52**	-	-		−0.52**	−0.32*	−0.35*
Emotions	-	−0.44**	−0.37*		-	-	-
Overall HRQOL	−0.3	−0.32*		−0.33*	−0.43**	−0.31*	−0.38*

**p < .01; *p < .05; one-tailed; Group with TBI, N = 31; IFG, Inferior Frontal Gyrus; CC, Cingulate Cortex; MPFC, Medial Prefrontal Cortex; Flexibility, Fluency & Working Memory,; S & S-R, Social & Self -Regulation; M & A, Motivation & Attention; DL&A, Daily Life & Autonomy.

Stepwise regression analysis revealed that the volume of the cingulate cortex could predict EF performance, F (1, 30) = 13.30, p = .001, adjusted R² = 0.29. Specifically, greater volume in the cingulate cortex predicted higher EF scores, B = 6.19, t (30) = 3.65, p = .001.

Further analyses of the EF-related regions’ predictive value on the DEX-R indices revealed greater temporal pole volume predicted greater fluency, flexibility, and working memory awareness, B = 63.78, t (30) = 3.30, p = .003, F (1, 30) = 10.88, p = .003, adjusted R² = 0.25, greater motivation and attention awareness, B = 43.64, t (30) = 3.57, p = .001, F (1, 30) = 12.73, p = .001, adjusted R² = 0.28, and better overall SA, B = 179.1., t (30) = 3.20, p = .003, F (1, 30) = 10.24, p = .003, adjusted R² = 0.24. In addition, when regressing ROIs’ volumes on SRSI indexes, less volume in the putamen, B = −23.43, t (30) = −2.78, p = .009, and the MPFC, B = −16.16, t (30) = −2.66, p = .009, predicted worse strategic awareness, F (1, 30) = 10.62, p = .0001, adjusted R² = 0.39. Furthermore, greater emergent/online SA deficits were predicted by reduced volume in the caudate, B = −24.09, t (30) = −3.21, p = .003.

For SA-related regions, stepwise regression indicated that greater atrophy in the temporal cortex predicted greater deficits in motivational and attentional SA, B = 57.54, t (30) = 2.95, p = .006, F (1, 30) = 8.68, p = .006, adjusted R² = 0.20. Furthermore, results showed that participants with TBI were less likely to be motivated to change given that they presented with less volume in the right globus pallidus, F (1, 30) = 9.53, p = .004, adjusted R² = 0.22, B = −109.70, t (30) = −3.09, p = .004. Also, strategic awareness was predicted by the cingulate cortex, F (1, 30) = 11.81, p = .002, adjusted R² = 0.27, B = −17.44, t (30) = −3.44, p = .002, indicating that reduced volume in the cingulate cortex leads to greater strategy – related SA deficits. Finally, greater impairment in emergent/online SA was predicted by less volume in the temporal pole, B = −15.25, t (30) = −3.17, p = .004, F (1, 30) = 10.05, p = .004, adjusted R² = 0.23.

Finally, stepwise regression revealed that SA-related areas could predict satisfaction with QOL and health-related QOL indexes. Specifically, greater volume in the temporal area predicted greater dissatisfaction with one’s psychological health, B = −54.07, t (30) = −3.22, p = .003, F (1, 30) = 10.39, p = .003, adjusted R² = 0.24, one’s perception of self, B = −266.09, t (30) = −3.96, p = .004, F (1, 30) = 15.66, p = .0001, adjusted R² = 0.33, and one’s overall health-related QOL, B = −29.97, t (30) = −3.27, p = .003, F (1, 30) = 10.72, p = .003, adjusted R² = 0.25.

Discussion

This study investigated the chronic and persistent course of brain volume loss in areas relating with EF, and SA; and whether volume loss in these regions reflected on EF and SA deficits, and as an effect in QOL.

Initially, overall GM and WM volumes were compared between the group with TBI and a neurotypical group, to validate the presence of volume loss in chronic TBI. Participants with TBI displayed less GM and WM volume, which was further coupled with greater CSF volume, as compared to controls. In addition, results from whole-brain analysis revealed atrophy in the left MPFC, the middle frontal gyrus, and the right cerebral WM. These findings are consistent with literature highlighting the persisting and long-term effects of chronic moderate-to-severe TBI on brain volume in these regions (7,10,11).

In addition, it was further sought to validate whether the abovementioned volume loss was reflected in neural systems implicated in the concepts of EF and SA. Significant volumetric group differences were detected for both EF and SA – related brain areas. Specifically, participants with TBI presented with reduced volume in EF-related areas including the globus pallidus, the OFC, the MPFC, the putamen, the temporal cortex, the right temporal pole, the left thalamus, the insula, the caudate, the right cingulate cortex, and the MPFC, compared to controls. Findings support past evidence by indicating that brain atrophy following a TBI is concentrated in a fronto-temporal network, the cerebellum, the hippocampus, and areas relating to the thalamic network, such as the insula, the caudate, the cingulate cortex, and the putamen (7).

The fronto-temporal network and underlying subcortical areas were further evident when investigating group differences in SA-related areas. Again, participants with TBI showed significantly less volume compared to the neurotypical group in areas similar to the EF-related regions, including the MPFC, OFC, the cingulate cortex, the temporal cortex and pole, the inferior frontal gyrus and the hippocampus, the insula and the putamen, and the globus pallidus. These findings highlight the similarities in the pathophysiologic sequelae between EF and SA – related regions, post-TBI. Therefore, it was further sought to investigate the interrelation between EF and SA -related areas. Results support an interrelation network between these areas, indicating that brain regions implicated in the EF and SA share common neurophysiology. This is the primary contribution of this study, further enhancing past theories arguing for a close association between the executive system and SA, highlighting the fact that for SA to be intact the EF should also be unimpaired (9). This finding may also explain the persistent SA deficits in cases with chronic TBI also presenting with atrophy and impairment in EF; as opposed to others, where SA returns to pre-injury levels.

The second contribution of this study was to examine whether neuropsychosocial group differences were reflected in the neural substrates involving EF and SA, for areas where the groups exhibited volumetric differences. Specifically, participants with TBI underperformed on the EF composite and presented with significant deficits in SA regarding attention and motivation, social and self – regulation, and overall SA. In addition, participants with TBI reported greater satisfaction with the environmental and greater dissatisfaction with the physical aspects of their QOL (for a detailed discussion of these findings, see 8).

These differences are reflected in both EF and SA - related cortical and subcortical regions with these regions showing significant predictive value for EF, SA, and health-related QOL. As expected, participants with TBI with less volume in the cingulate cortex were more likely to underperform in EF. This finding is consistent with evidence supporting that greater brain volume in EF-related regions, including subcortical areas relates with better neuropsychological performance (7), and specifically executive functioning (23).

Furthermore, the predictive value of both the EF and SA – related areas on measures of SA was investigated, with volume in different regions predicting different aspects of SA. Specifically, findings support the implication of reduced temporal pole volume in greater deficits in fluency, flexibility, and working memory, and overall SA. In addition, greater atrophy in both the temporal pole and cortex predicted greater deficits in motivational and attentional SA. Less volume in the putamen, cingulate cortex, and the MPFC predicted worse strategic awareness. Finally, greater impairment in emergent awareness was predicted by reduced volume in the temporal pole and the caudate. These findings highlight the implication of EF and SA - related brain areas in SA, further enhancing the association between these two functions, the implication of a specific network involved in SA (see 24), and how damage to the underlying neurocircuit of these systems may be involved into anosognosia.

Further supporting the aforementioned arguments was evidence showing that greater volume in SA-related structures, such as the right globus pallidus, was predictive of reduced motivation to change post-injury disability. The latter finding may be informative of the concept of SA as a whole, as individuals with greater volume and fewer SA deficits are more likely to be aware of the effort required to engage in altering dysfunctional behaviors (33,34).

The final contribution of this study was the predictive value of EF and SA – related brain structures in the subjective experience of participants with TBI regarding their QOL. Participants with TBI presenting with reduced volume in the temporal area were more likely to report greater satisfaction with their overall health-related QOL. In addition, greater atrophy in the temporal area reflected on both their QOL and health-related QOL, with participants with TBI reporting higher levels of satisfaction regarding their psychological health and their sense of self, respectively. This evidence lends support to findings that participants with TBI presenting with executive dysfunction and SA deficits are more likely to report greater satisfaction with different aspects of their lives (8,17). These findings are novel, as no evidence exists to indicate the implication of brain atrophy on QOL/health-related QOL following chronic TBI, and shed light to the perplexed relationship between EF, SA and QOL.

Findings from this study depict significant brain atrophy and its related effects lingering for several years post-TBI. This study allowed capturing the true post-injury effects, due to the fact that post-acute rehabilitation services offered in Cyprus are limited, highlighting the importance of systematic post-acute comprehensive rehabilitation and community re-integration. Furthermore, the interrelation of EF and SA – related brain regions along with findings regarding executive functioning and metacognition with the subjective experience of QOL may lead to the development of biopsychosocial models or the enhancement of existing models discussing these associations. Therefore, such evidence linking impaired SA to EF and to QOL in participants with TBI not having received systematic post-acute rehabilitation may guide health-care professionals in designing more comprehensive acute and post-acute rehabilitation programs focusing on the impact of impairments in EF and SA on daily participation and QOL. Finally, these results may inform on the significance of more appropriate and well-equipped rehabilitation centers specific to neurological conditions, including TBI.

Despite the novelty of these findings, further investigation is required on how these constructs act as moderators to recovery using larger samples. Furthermore, this study attempted to preserve a homogeneous sample by recruiting male participants with chronic moderate-to-severe TBI, alone. Hence, future studies should replicate these analyses using female participants, which may allow detecting potential variance in findings due to gender differences (e.g. hormones, see 35); and in individuals with less severe injuries such as mild TBI, or of greater chronic course, that is, greater than 6 years, or during the acute phase. Therefore, further exploration of these effects should be investigated on a severity and TSI continuum. Also, it is important to highlight the need for longitudinal studies and repeated assessments in populations with chronic conditions such as TBI, in order to keep track of potential neuropsychological and psychosocial changes, as well as further brain atrophy.

Acknowledgments

The authors wish to thank the Nicosia General Hospital Intensive Care Unit, and the Melathron Agoniston EOKA for the referrals, and all participants and their families for their participation.

Disclosure statement

No potential conflict of interest was reported by the author(s).

Supplementary material

Supplemental data for this article can be accessed on the publisher’s website

Additional information

Funding

This work was supported by the Cyprus Research Innovation Foundation and co-funded by the Republic of Cyprus and the European Regional Development Fund under the Framework Programme “RESTART 2016-2020/EXCELLENCE HUBS: PILLAR II” for Research, Technological Development and Innovation (EXCELLENCE/1216/0411) and a new infrastructure grant (NEW INFRASTRUCTURE/STRATEGIC/0309/37);Research and Innovation Foundation [EXCELLENCE/1216/0411, NEW INFRASTRUCTURE/STRATEGIC/0309/37];