INTRODUCTION
White matter abnormalities have been observed in children and adolescents born very preterm (VPT: Nosarti et al., Reference Nosarti, Giouroukou, Healy, Rifkin, Walshe and Reichenberg2008). The corpus callosum (CC) is the largest white matter structure in the brain, connecting the hemispheres and underpinning inter-hemispheric processing. Reductions in mid-posterior and posterior CC size have been described in VPT children and adolescents, that is, <33 weeks gestation (Nosarti et al., Reference Nosarti, Rushe, Woodruff, Stewart, Rifkin and Murray2004; Nosarti et al., Reference Nosarti, Giouroukou, Healy, Rifkin, Walshe and Reichenberg2008). These differences may be associated with gestational age, for instance, posterior CC size is reduced in VPT adolescents born at 28–30 weeks, with those born at 31–33 weeks spared (Narberhaus et al., Reference Narberhaus, Segarra, Caldu, Gimenez, Junque and Pueyo2007).
Studies that follow VPT individuals into adulthood are limited, and it is not clear whether differences in CC size persist, or represent developmental delay. For instance, a 3.3% CC area increase was observed in healthy controls between 15 and 19 years versus 13.4% in VPT participants (Allin et al., Reference Allin, Nosarti, Narberhaus, Walshe, Frearson and Kalpakidou2007). Nevertheless, white matter abnormalities have been observed in the genu of CC in VPT adults (Kontis et al., Reference Kontis, Canti, Cuddy, Walshe, Nosarti and Jones2009), and functional and structural differences persist in other brain regions in VPT adults (Lawrence et al., Reference Lawrence, Rubia, Murray, McGuire, Walshe and Allin2009, Reference Lawrence, McGuire, Allin, Walshe, Giampietro and Murray2010; Narberhaus et al., Reference Narberhaus, Lawrence, Allin, Walshe, McGuire and Rifkin2009; Nosarti et al., Reference Nosarti, Giouroukou, Healy, Rifkin, Walshe and Reichenberg2008).
Empathy is a multi-dimensional construct with cognitive and affective components. Cognitive empathy is to understand what another individual is thinking/feeling, and affective empathy is an emotional response to the mental state of another, excluding self-orientated emotions such as personal distress (Baron-Cohen & Wheelwright, Reference Baron-Cohen and Wheelwright2004; Davis, Reference Davis1980). Empathy deficits have been observed in individuals with agenesis of the corpus callosum (Paul et al., Reference Paul, Brown, Adolphs, Tyszka, Richards and Mukherjee2007), and aberrations in corpus callosum morphology have been found in cohorts with known empathy disturbances such as autistic spectrum disorders (ASDs; Chung et al., Reference Chung, Dalton, Alexander and Davidson2004). For example, Chung et al. (Reference Chung, Dalton, Alexander and Davidson2004) found reductions in anterior regions (genu and rostrum) alongside reduced white matter density in posterior CC regions in participants diagnosed with ASDs. Whether the decreases in posterior CC area are associated with an empathy deficit in VPT populations has yet to be examined, although increased shyness and low extroversion have been observed in VPT and low birth weight adults (Allin et al., Reference Allin, Rooney, Cuddy, Wyatt, Walshe and Rifkin2006; Schmidt et al., Reference Schmidt, Miskovic, Boyle and Saigal2008).
The aim of this study was to investigate empathy and CC size in VPT adults and controls using standardized self-report measures and high-resolution structural magnetic resonance imaging (MRI). We administered two empathy questionnaires: the Interpersonal Reactivity Index [IRI] (Davis, Reference Davis1980) and the Empathy Quotient [EQ] (Baron-Cohen & Wheelwright, Reference Baron-Cohen and Wheelwright2004). We predicted reduced posterior CC size in VPT adults in comparison to controls, and tested whether such differences were associated with deficits in self-reported empathy.
METHOD
Participants
In 1983–1984, 252 infants born at <33 weeks gestation were admitted within 5 days of birth to the Neonatal Unit at University College London Hospital, survived, and were discharged. All individuals born at <=28 weeks gestation were enrolled for long-term follow-up, as well as a random sample of those born from 29 to 33 weeks gestation. One hundred forty-seven (40% of sample) adolescents were selected for study (78 born at <28 weeks and 69 born at 29–33 weeks gestation). One hundred thirteen individuals were assessed in adolescence; 57 were born at <=28 weeks and 56 from 29 to 33 weeks gestation. In 2003, participants who met the inclusion criteria, i.e., no history of cerebral palsy, grade 3/4 intraventricular hemorrhage, or cystic periventricular leucomalacia, were contacted. Those who responded took part in a series of studies (Lawrence et al., Reference Lawrence, Rubia, Murray, McGuire, Walshe and Allin2009, Reference Lawrence, McGuire, Allin, Walshe, Giampietro and Murray2010; Narberhaus et al., Reference Narberhaus, Lawrence, Allin, Walshe, McGuire and Rifkin2009). Control participants were also recruited via advertisements in the local press and universities. Inclusion criteria were full-term birth (37–42 weeks gestation) and English as a first language; exclusion criteria included birth complications (e.g., birth weight <2500 g, endotracheal mechanical ventilation), prolonged gestation (>42 weeks), history of psychiatric illness, severe hearing and motor impairment.
Twenty-six participants took part in the current study. Ethical approval was obtained from the local committee, and written informed consent was obtained. There were 9 controls (5 male; mean age of 18.9 years ± 1.1) and 17 VPTFootnote 1 participants (9 male: mean age of 19.9 ± .8). Mean gestational age for the VPT group was 28.4 (± 2.1) weeks, and mean birth weight was 1212.9 grams (± 3020), and neither statistically differed from the larger cohort from whom they were drawn (gestational age: t = .78; p > .05; birth weight: t = .5; p > .05). A total of 35.3% of the sample had an Apgar score between 0 and 3 at 1 min, 35.3% were rated between 4 and 6, and 29.4% scored between 7–10 which is within normal range. These scores did not differ from the distribution in the larger cohort (X(2) = .1.2; p > .05). There was no difference between the cohorts in the proportion of multiple births (X(5) = 3.5; p > .05).
Measures
The IRI (Davis, Reference Davis1980) and EQ (Baron-Cohen & Wheelwright, Reference Baron-Cohen and Wheelwright2004) were used to assess empathic ability. These gold standard measures tap slightly different aspects of empathy. The EQ contains 40 empathy items, with ratings on a 4-point scale, ranging from, strongly agree to strongly disagree. A score of 0 is allocated for a nonempathic response and 1–2 for an empathic response. The scale was designed to be unifactorial; however, principal components analysis found 13 items correspond to cognitive empathy, including statements such as “I can tell if someone is masking their true emotion with a maximum score of 26”; 9 relate to emotional reactivity such as “seeing people cry doesn’t really upset me” with a maximum score of 18; 6 questions relating to social skills, that is, “I often find it difficult to judge whether something is rude or polite” with a maximum score of 12 (Lawrence et al., Reference Lawrence, Shaw, Baker, Baron-Cohen and David2004). The current study results were analyzed using this 28-item solution.
The IRI contains three empathy subscales each with seven items rated on a 5-point scale ranging from “does not describe me at all” to “describes me very well,” with, a maximum score of 35. The subscales measure perspective taking, e.g., “I sometimes find it difficult to see things from the other guy’s point of view”; empathic concern, e.g., “I often have tender, concerned feelings for people less fortunate than me”; and personal distress, nonempathic self-orientated emotional contagion, e.g., ‘“Being in a tense emotional situation scares me.”.
Procedure
A 1.5 Tesla (T) GE Signa Horizon MRI scanner was used to obtain sagittal T2-weighted fast spin-echo, 27 × 4 mm contiguous slices; axial T2-weighted double-echo fast spin-echo, 28 × 5 mm contiguous slices and three-dimensional T1-weighted gradient-echo sequence that allowed reconstruction in any plane of one hundred twenty-four 1.5-mm slices. CC area was measured on a mid-sagittal slice by sub-dividing it into four regions by drawing lines perpendicular to its anterior–posterior length using image analysis software Analyze (Nosarti et al., Reference Nosarti, Rushe, Woodruff, Stewart, Rifkin and Murray2004). Inter-rater reliability was performed on five randomly selected CC area measurements from three independent raters (anterior α = .98; p < .001; mid-anterior α = .99; p < .001; mid posterior α = .99; p < .001; posterior α = .99; p < .001). Total white matter volume was extracted using Voxel-based Morphometry in SPM2.
RESULTS
There were six instances of missing data which were replaced with the mean of the sample, except for data from one control participant who failed to complete six personal distress subscale items and so was excluded from this analysis.
Gender did not differ between groups (X 2 = .016; df 1; p > .05); however, VPT and control participants did differ in age (t (24) = 2.8; p = .01), likely due to the narrow age range in the VPT group. Mean Full-Scale IQ estimated by the Wechsler Abbreviated Scale of IntelligenceFootnote 2 was 106 (± 12) for controls (n = 7) and 98.9 (± 10.4) for VPT participants, with no between-group difference (t (22) = 1.5; p > .05).
There was no significant between-group difference in total white matter volume (t (24) = 1.1; p > .05). However, total white matter volume was positively associated with total CC area across groups (r = .53; p = .003), so the data were analyzed adjusting for total white matter volume and age. There were no significant differences between controls and VPT adults in total CC area (F (1,22) = .63; p > .05), anterior CC area (F (1,22) = .3; p > .05), mid-anterior CC area (F (1,22) = .03; p > .05), nor mid-posterior CC area (F (1,22) = .51; p > .05). However, posterior CC area was reduced in VPT adults (F (1,22) = 4.9; p = .04). In addition, white matter volume more closely related to posterior CC area in controls, than VPT adults (F (1,22) = 4.6; p = .04).
To examine the influence of gestational age on CC area, the sample was divided into three groups (Lawrence et al., Reference Lawrence, McGuire, Allin, Walshe, Giampietro and Murray2010); 1) =< 28 weeks gestation (n = 7), 29–32 weeks gestation (n = 10), 3) controls => 37 weeks gestation (n = 9) – see Figure 1. A significant difference was also observed in anterior CC area between the two VPT groups (F (2,21) = 3.8; p = .04). No between-group differences were observed in the mid-anterior region (F (2,21) = 2; p > .05). A near significant positive linear trend was observed for mid-posterior CC area with those with the lowest gestational age having the largest reduction (F (2,21) = 2.7; p = .09). A positive linear trend was also observed across groups for posterior CC area (F (2,21) = 4.5; p = .02). Again, white matter volume was more closely related to posterior CC area in controls than VPT participants (F (2,21) = 5.4; p = .03).
Mean Corpus Callosum Area in mm2 by Gestational Age and Adjusted for Age and White Matter Volume. Error bars represent SE of Mean
Empathy
To investigate between-group empathy differences, univariate analyses of variance (ANOVAs) were conducted controlling for age. VPT adults had higher scores on the personal distress subscale of the IRI (F (1,22) = 4.6; p = .04). No other significant group differences were observed (Table 1). No linear trends in self-reported empathy were observed when dividing participants by gestational age including for personal distress (F (2,21) = 2.19; p = .12). Kendall’s partial correlation analysis adjusted for age revealed a statistical trend indicative of a negative association between personal distress and anterior CC size in VPT (r = −.38; p = .07), but not control participants (r = −.05; p = .46).
Mean and SD Scores on EQ and IRI Subscales Stratified by Group
Note
EQ = Empathy Quotient; IRI = Interpersonal Reactivity Index.
DISCUSSION
With total white matter volume adjusted, VPT adults show reduced posterior CC area in contrast to controls. A positive linear trend between posterior CC area and gestational age was also observed, with similar findings for mid-posterior CC area. Participants who were born at <= 28 weeks gestation also had reduced anterior CC size in contrast to VPT adults born nearer term. No significant differences were observed between VPT adults and controls in self-reported empathy, although the VPT group reported increased personal distress in response to social stimuli, in contrast to controls. Correlational analyses suggest this effect may be related to reductions in anterior CC area.
That VPT participants had reduced posterior CC area provides further evidence that structural CC differences observed in younger cohorts persist into adulthood (Caldu et al., Reference Caldu, Narberhaus, Junque, Gimenez, Vendrell and Bargallo2006; Narberhaus et al., Reference Narberhaus, Segarra, Caldu, Gimenez, Junque and Pueyo2007, Reference Narberhaus, Segarra, Caldu, Gimenez, Pueyo and Botet2008; Nosarti et al., Reference Nosarti, Rushe, Woodruff, Stewart, Rifkin and Murray2004). We also observed a positive linear trend between gestational age and posterior CC area, and mid-posterior CC area (p = .09), suggesting that reductions in CC area may be more severe in those with younger gestational age. CC injury may be partly explained by the vulnerability of the developing corpus callosum to hypoxic–ischemic damage and hemorrhage, possibly due to the intrinsic vulnerability of immature oligodendrocytes. Previous studies have observed an association between CC size and gestational age in VPT adolescents (Caldu et al., Reference Caldu, Narberhaus, Junque, Gimenez, Vendrell and Bargallo2006) and the splenium, which encompasses the posterior section of the CC, is only reduced in VPT adolescents with gestational age <= 30 weeks (Narberhaus et al., Reference Narberhaus, Segarra, Caldu, Gimenez, Junque and Pueyo2007).
We also observed reductions in anterior CC area in VPT participants of <28 weeks gestation in contrast to those of 29–32 weeks gestation. This is consistent with VPT adolescent data which suggests those with gestational age <27 weeks, to have significantly reduced anterior CC area in contrast to controls (Narberhaus et al., Reference Narberhaus, Segarra, Caldu, Gimenez, Junque and Pueyo2007). Although in the current study, participants of 29–32 weeks gestation appeared to have slightly larger anterior CC size than controls, this difference was not statistically significant.
VPT adults did not self-report empathy deficits, and scores for all groups were consistent with normative data for these questionnaires (Lawrence et al., Reference Lawrence, Shaw, Baker, Baron-Cohen and David2004; Lawrence et al, Reference Lawrence, Shaw, Giampietro, Surguladze, Brammer and David2006). This suggests differences in CC area may be unrelated or at least not crucial to empathic processing. Previous studies suggest that, although agenesis of the CC leads to high psychiatric morbidity, this is unrelated to any specific pathology (Taylor & David, Reference Taylor and David1998). However, it is also possible that this finding is due to the processes of cognitive or neural plasticity, for instance, VPT adults have been previously found to recruit additional brain structures for successful task completion (Lawrence et al., Reference Lawrence, Rubia, Murray, McGuire, Walshe and Allin2009, Reference Lawrence, McGuire, Allin, Walshe, Giampietro and Murray2010).
VPT adults did, however, report higher rates of personal distress than controls. Personal distress is an emotional reaction to social situations, which is self-orientated. Although not an empathic response, it is a form of emotional contagion which includes feelings of unease and distress (Lawrence et al., Reference Lawrence, Shaw, Giampietro, Surguladze, Brammer and David2006). These data are consistent with other personality differences in VPT adults such as elevated anxiety (Allin et al., Reference Allin, Rooney, Cuddy, Wyatt, Walshe and Rifkin2006), increased shyness, and reduced extroversion (Schmidt et al., Reference Schmidt, Miskovic, Boyle and Saigal2008).
A statistical trend indicated that differences in anterior CC size may be modestly associated with personal distress in VPT adults, with larger anterior CC reductions associated with increased personal distress. Previous studies have found structural differences in the anterior CC, which contains interhemispheric fibers connecting different sections of the prefrontal cortex, to be associated with prefrontal functions in VPT adolescents (Narberhaus et al., Reference Narberhaus, Segarra, Caldu, Gimenez, Pueyo and Botet2008). It has been suggested that personal distress may result from a lack of downregulation of a vicarious affective response (Lawrence et al., Reference Lawrence, Shaw, Giampietro, Surguladze, Brammer and David2006), a skill requiring prefrontal inhibition mechanisms. Consistently, the anterior CC is larger in participants with an enhanced ability to downregulate pain (Horton et al., Reference Horton, Crawford, Harrington and Hunter-Downs2004). This account may explain the association observed between personal distress and anterior CC size in VPT adults. Future studies are needed to determine whether differences in emotion regulation in VPT adults of low gestational age underpin observed personality differences (Allin et al., Reference Allin, Rooney, Cuddy, Wyatt, Walshe and Rifkin2006; Schmidt et al., Reference Schmidt, Miskovic, Boyle and Saigal2008).
Overall, these data are limited by being reliant on self-report. The small sample size and lack of statistical power may also have prevented the observation of between-group empathy differences, as response rate for controls for was lower than for the VPT group. However, this is unlikely as the reported empathy scores were consistent with normative data. As hypothesized, VPT adults had significantly smaller posterior CC area than controls, which was related to gestational age. In addition, VPT participants with a gestational age <=28 weeks had reduced anterior CC size, in contrast to VPT adults born nearer term, and this may be related to their increased levels of personal distress. Further studies with a larger sample size are warranted, and it would be prudent to use recent advances in neuroimaging methods such as Diffusion Tensor Imaging and connectivity analyses.
ACKNOWLEDGMENTS
We thank all the participants for giving their time and Anastasia Kalpakidou for her help with data collection. We can confirm that there are no potential or real conflicts of interest. This work was supported by grants awarded by the March of Dimes Birth Defects Foundation, USA (12-FY03-4); and the Health Foundation, UK (1206/2063) to RMM, CN, PKM and LR.
