Bipolar disorder (BD) is a severe psychiatric disorder with a prevalence of 1–2%, with recurring episodes varying from psychosis to mania or major depression. It has a deep social impact due to increased suicide risk and poor quality of life, and is often associated with disability and chronicity, especially if there is a delay in treatment (Altamura et al. Reference Altamura, Dell'Osso, Berlin, Buoli, Bassetti and Mundo2010, Reference Altamura, Buoli, Caldiroli, Caron, Cumerlato Melter, Dobrea, Cigliobianco and Zanelli Quarantini2015). Many neuroimaging studies demonstrated brain abnormalities in patients affected by BD, afflicting both white and gray matter (Bellani et al. Reference Bellani, Boschello, Delvecchio, Dusi, Altamura, Ruggeri and Brambilla2016; Maggioni et al. Reference Maggioni, Bellani, Altamura and Brambilla2016). In particular, the inter-hemispheric connectivity, primarily fronto-limbic and callosal connectivity, results to be disrupted (Brambilla et al. Reference Brambilla, Bellani, Yeh, Soares and Tansella2009; Sprooten et al. Reference Sprooten, Barrett, McKay, Knowles, Mathias, Winkler, Brumbaugh, Landau, Cyr, Kochunov and Glahn2016) and subcortical abnormalities have also been recently reported (Hibar et al. Reference Hibar, Westlye, van Erp, Rasmussen, Leonardo, Faskowitz, Haukvik, Hartberg, Doan, Agartz, Dale, Gruber, Krämer, Trost, Liberg, Abé, Ekman, Ingvar, Landén, Fears, Freimer, Bearden, Sprooten, Glahn, Pearlson, Emsell, Kenney, Scanlon, McDonald, Cannon, Almeida, Versace, Caseras, Lawrence, Phillips, Dima, Delvecchio, Frangou, Satterthwaite, Wolf, Houenou, Henry, Malt, Bøen, Elvsåshagen, Young, Lloyd, Goodwin, Mackay, Bourne, Bilderbeck, Abramovic, Boks, van Haren, Ophoff, Kahn, Bauer, Pfennig, Alda, Hajek, Mwangi, Soares, Nickson, Dimitrova, Sussmann, Hagenaars, Whalley, McIntosh, Thompson and Andreassen2016). Furthermore, gray matter thickness and volume are heavily affected (Houenou et al. Reference Houenou, d'Albis, Vederine, Henry, Leboyer and Wessa2012; Hanford et al. Reference Hanford, Nazarov, Hall and Sassi2016), especially in prefronto-temporal areas.

The pathogenesis of this disease is not entirely clear yet, but there is evidence of a predominant contribution of genetic factors to the risk for BD, with a very high heritability estimated around 85% (McGuffin et al. Reference McGuffin, Rijsdijk, Andrew, Sham, Katz and Cardno2003). Classical genetic research involving families, twins and also adoptions showed that genes are strictly related to the risk of developing BD (Craddock & Sklar, Reference Craddock and Sklar2013). Over the last years, the study of twins has proven to be particularly effective in biomedical etiological research. Monozygotic (MZ) twins are genetically identical and dizygotic (DZ) twins share 50% of their genes; also, both MZ and DZ twins share environmental factors in utero as well as within the family in early infancy. For these reasons, twin studies allow etiologic inferences to be made without the confounding effect of unmeasurable factors such as genetic background, intrauterine or perinatal exposures, or family environment. Thus, the main challenges associated with case-control studies are overcome when dealing with twins (McGue et al. Reference McGue, Osler and Christensen2010).

Discordant-twin studies, in particular, could be crucial for the understanding of the interplay between these factors (Fagnani et al. Reference Fagnani, Bellani, Soares, Stazi and Brambilla2014). In this review, we address twin studies, which focus on magnetic resonance imaging (MRI) of the brain and BD. Ten studies met our inclusion criteria (i.e., MR imaging, twin pairs affected by BD, comparison with healthy twin pairs, focus of work on genetic influence on BD). Their main findings are summarised in Table 1.

Table 1. Selection of twin studies on BD investigating brain structure and function with MRI

MZ, monozygotic; DZ, dizygotic; BD, bipolar disorder; HC, healthy controls; SCZ, schizophrenia; SEM, structural equation modelling; WM, white matter; GM, gray matter; MRI, magnetic resonance imaging; ROI, region of interest.

Based on the assumption that environmental factors are shared by MZ and DZ twins to the same extent (‘Equal Environments Assumption’) (Neale & Cardon, Reference Neale and Cardon1992), a higher similarity observed in MZ twins suggests genetic influences on the trait under study. Consequently, the comparison of BD-affected twins with healthy control (HC) twins could help in shedding light on the mechanisms of the disease. Noga et al. (Reference Noga, Vladar and Torrey2001) compared a small sample (6 pairs) of discordant MZ twins to MZ HC and showed that left caudate was larger in BD and co-twins, suggesting genetic effects, and right caudate was larger only in BD, implying environmental factors. Kieseppä et al., in two works (Reference Kieseppä, van Erp, Haukka, Partonen, Cannon, Poutanen, Kapri and Lönnqvist2002, Reference Kieseppä, van Erp, Haukka, Partonen, Cannon, Poutanen, Kaprio and Lönnqvist2003), found evidence of genetically-induced decreased left white matter and environment-related decreased frontal white matter, while no significant results were found for gray matter, in a dataset comprising around 30 BD-affected twins. Bearden et al. (Reference Bearden, van Erp, Dutton, Boyle, Madsen, Luders, Kieseppa, Tuulio-Henriksson, Huttunen, Partonen, Kaprio, Lönnqvist, Thompson and Cannon2011) focused on the white matter and found callosal thinning, area reduction and different ventral curvature in patients with BD (n = 21) compared with both co-twins (n = 19) and controls (n = 34), while co-twins had no differences with controls. This suggests that differences in corpus callosum are disease- rather than genetically-induced.

The only study, which analysed fMRI task activation found no difference in BD twins or their co-twins, in respect to controls during word generation, while it found relevant results in schizophrenia (Costafreda et al. Reference Costafreda, Fu, Picchioni, Kane, McDonald, Prata, Kalidindi, Walshe, Curtis, Bramon, Kravariti, Marshall, Toulopoulou, Barker, David, Brammer, Murray and McGuire2009).

A more formal description of the genetic and environmental estimates that can be obtained with the classical twin model involves three factors, namely additive genetic (A), common (i.e., shared by twins) environmental (C) and unique (i.e., individual-specific) environmental (E) factors, under the so-called ACE model. This model allows one to partition the total variance in liability to a given disease (e.g., BD) in the three components A, C and E; in particular, the proportion of total variance due to the A component is named ‘heritability’. Such a decomposition requires structural equation model (SEM) fitting, which has limited applications in clinical contexts due to the large sample size needed to achieve adequate statistical power (Wolf et al. Reference Wolf, Harrington, Clark and Miller2013). The SEM approach has been employed in five of the studies considered in this review, and all of them found no significant role of common environment.

Considering a population of around 200 twins at baseline (MZ and DZ, both discordant and concordant for BD, details in Table 1) and 100 twins at follow-up, Bootsman et al. (Reference Bootsman, Brouwer, Kemner, Schnack, van der Schot, Vonk, Hillegers, Boomsma, Hulshoff Pol, Nolen, Kahn and van Haren2015) found a phenotypic and genetic association of BD with smaller subcortical volumes at baseline. Volume change over time had low heritability, but high association with unique environment. Interestingly, most of the other studies, which used the ACE model found only environmental influences on gray matter, while they detected genetic effects on white matter. Van der Schot et al. (Reference van der Schot, Vonk, Brouwer, van Baal, Brans, van Haren, Schnack, Boomsma, Nolen, Hulshoff Pol and Kahn2010) demonstrated, in a sample of around 200 individuals (49 affected twin pairs and 67 healthy twin pairs), that genetic factors are involved in white matter density of superior longitudinal fasciculus, while cortical gray matter volume decrease was related only to unique environmental factors. Another work of the same group (Van der Schot et al. Reference van der Schot, Vonk, Brans, van Haren, Koolschijn, Nuboer, Schnack, van Baal, Boomsma, Nolen, Hulshoff Pol and Kahn2009) showed that the genetic risk of developing BD was associated with decreases in white matter volumes, while gray matter was highly related to unique environment. This indicates that genes involved in BD could contribute to white matter loss found in BD patients and their co-twins, while gray matter decrease is probably related to the illness itself. Hulshoff et al. (Reference Hulshoff Pol, van Baal, Schnack, Brans, van der Schot, Brouwer, van Haren, Lepage, Collins, Evans, Boomsma, Nolen and Kahn2012) found relevant genetic factors in both BD and SCZ, related to smaller white matter volume and thickness: thinner in parahippocampus and right orbitofrontal cortex, thicker in temporoparietal and left superior motor cortices. In this case, gray matter was influenced, although not significantly, by environmental factors. Vonk et al. (Reference Vonk, van der Schot, van Baal, van Oel, Nolen and Kahn2014) found an indirect genetic relationship between the genetic risk of developing BD and brain, white matter and cortical volumes: these quantities were genetically related with the dermatoglyphic-derived ridge count, and this was related with the risk of BD.

In summary, twin studies demonstrated that there are strong genetic factors involved in the pathogenesis of BD, which also influence white matter, which in turn is involved in brain connectivity. Interestingly, corpus callosum seems to be disease-related. Gray matter, on the contrary, seems more affected by environmental effects or by the disease itself. These results have been found employing different methodologies (VBM, ROI-based studies, automatic estimation of cortical thickness and brain volumes), which allow the study of brain morphology from many points of view. It would be beneficial to introduce more recent techniques as cortical folding or gyrification also in twin studies, to reach a more comprehensive understanding of brain characteristics. A future use of the twin design should be encouraged, especially exploiting the potential of population-based Twin Registries, which could help in the identification of high numbers of BD concordant and discordant pairs, thus facilitating the complex modelling of the genetic and environmental etiological mechanisms.