We identified three overall sparsely connected states, and one state with strong positive and negative couplings. Although all groups presented similar connectivity states, TM and CW presented more instability in the temporal meta-states transitions than CM; and TM spent more time in the more densely connected state than CM. In addition, TM had stronger connectivity than CM in the coupling between the visuospatial network node and the seed in the supplementary motor area of the salience. Finally, the attention regulation dimension of interoceptive awareness was positively associated with both states and meta-states metrics in TM, while such relation was not found in the cisgender groups.
The compelling finding was that TM dwelled more time in the moderately frequent densely connected state than CM. Remaining significantly more time in this state than in the other sparser state could be related to an increased saliency. The relevant role of the salience network has been previously stressed in terms of explaining gender phenotypes 4. Moreover, meta-states fluidity and range dynamics differentiated CM from the TM and CW. Contrary to recent cisgender works 20, CM had less meta-state dynamism than the two biologically females groups (TM and CW). In our previous stationary FC results 4 using the same sample 28, we reported decreased functional connectivity in TM with respect to CM. We could speculate that TM possesses increased integration large-scale properties in expenses of reduced segregation states than CM, resulting in an increased averaged connectivity of CM comparing them with TM. However, stationary graph-theoretical results failed to differentiate these two groups.
Interestingly, intergroup differences of the State’s 2 FC revealed that TM had increased connectivity between two nodes: the right supramarginal and the left supplementary motor area with respect to CM. We previously reported decreased connectivity of the supramarginal gyrus in TM with respect to CM 4; and thicker cortex in TM and CW with respect to CM using a different cohort 5. Functional connectivity differences in the supramarginal gyrus between cisgender 33 and transgender groups 11 were also reported in relation to the own body perception 11,33; and to interpersonal emotion processing 34. Multimodal translational approaches 35 are needed for understanding these compelling yet challenging findings contributing to the formation of the gendered-self.
We also described one discrete functional configuration with a coupling between the bilateral precentral gyrus of the sensorimotor network. We could associate this sensorimotor-state with the self-representation of the typical traits assigned at birth as suggested by the functional hypothesis 10. Indeed, increased activity in males in the precentral cortex in a small cohort of cisgender individuals has been described from a task-based self-body perception task 33. However, this state was slightly prevalent (13%), and TM’s connectivity did not differentiate from the cisgender groups.
One key element is present between our previous functional work 4 and the present: the salience network and its interactions. We could describe the salience as a switching network 36 that interplays mainly but not restricted to the default mode, executive control and itself, as well as the auditory and language networks. Interestingly, when defining the mind-wandering processes, the “general salience network” is associated with an automatic bottom-up salience detection that together with the core of the default exerts automatic constraints on thought; while the executive control exerts deliberate constraints by reducing those automatic constraints controlled by the other two networks 37.
In line with the frontal predominance described from the static seed-based connectivity approach 4, we also observed a frontal involvement between the FC couplings within states. More interestingly, two regions stood out: the insula and the paracingulate. In the transgender literature, projections between the pregenual anterior cingulate cortex and the insula 38 are related to the own body perception, as well as increased activity in the paracingulate when viewing bodies opposed to the sex at birth 11 in TM 10. The insula is described as a multimodal convergence area implicated in body awareness 39, and an integral hub 1,36.
In addition, we found a positive association between the spatiotemporal dynamic metrics and the attention regulation proxy of the interoceptive awareness scale only in the TM group. Interestingly, those TM who were more apt to switch connectivity states and range over greater meta-states spaces presented higher interoceptive awareness. We could speculate that the densely connected state (State 3) pertained to a sustained attention regulation to body sensations in relation to the experience of the resting-state acquisition in TM. State’s 3 FC was represented mainly by positive and negative couplings within the triple-network 1. Regionally, mind-wandering processes are associated with the default mode activity, while the change to an awareness of such mind-wandering periods is related to the interplay with the salience 37. Indeed, it is hypothesized that the salience network “increases the stability of attention over time by constraining the spontaneous movement of attention” 37.
Among the scarce literature on interoceptive awareness and brain imaging 40–42, a recent study reported self-regulated sustained attention by modulating the interaction between the salience and the default mode networks 43. Such studies reinforce the approach that mind, cognition and emotion emerge from the reciprocal interactions of the brain-body signals-external world. In TM such body self-identification requires of additional involvement of emotional processing 11,44.
Among the available hypotheses to explain gender, some try to answer the question on why the gender is built in the brain and others on how the brain works in relation to gender. Within the framework of the why, we proposed a neurodevelopmental cortical hypothesis 8 by means of structural MRI techniques in transgender people before undergoing hormone affirming treatment 5–7, that was later supported by brain functional MRI data 4. Based on our previous findings, TM have a mixture of masculine, feminine and defeminized morphological brain traits. Interestingly, cortical thickness of TM differentiated from CM (but not from CW) in a different pattern than that observed when comparing CM and CW. These patterns were also found when using functional data 4. When shifting the focus on the how, a predominant hypothesis suggests that gender incongruence could rest on the functional disconnection of fronto-parietal regions of the own-body self-referential network, also reporting a cortical thickening of the mesial prefrontal and precuneus cortices 10,25,38.
In this work, we can observe a similar pattern as we previously described. We previously suggested that sex differences in gender groups would be measure-dependent 8, conferring each gender a particular phenotype. Regionally, our temporal perspective agrees with the involvement of a sensorimotor-state that could be implicated in the so-called own-body perception network 10,33, and we postulate that the insula and its projections are not only implicated in own-body perception processes in transgender people 38, but as one of the brain regions that constitutes the core in building a gendered-self.
This novel whole-brain chronnectome technique suggests that the underlying brain mechanisms accounting for gender differences are not likely explained by interactions of one sole network, or from a functional disconnection of specific brain regions. We could hypothesize that the potential FC signatures underlying gender variability rely on a shared dynamical core that orchestrates differently, and in a gender fashion, the functional activity of the rest of the brain. Undoubtedly, disentangling the brain dynamics differences underlying the conceptualization of the gendered-self have a great potential as new dynamical models emerge 45–47. These models offer a causal mechanistic explanation to the brain dynamics constrained by functional and structural connectivity 45, that allow to study the information processing across time, space, and scales 46.
In this work, one important shortcoming is the TR > 2s, hindering a thorough assessment of the temporal brain dynamics which can be solved by multiband acquisition protocols. Secondly, transgender variants are relatively rare 48 and therefore, our sample size limited the robustness of the associations with the awareness questionnaire that have no global score.
In sum, we found that FC dynamics underscore the importance of the salience as keystone brain network and its interplay with other networks. In this process, the insula is one relevant region that makes possible the switch between the internal and the external world. We stress the importance of the study of whole-brain network interactions rather than regional connectivity differences of a priori seed-based analysis. Finally, we provide a functional explanation from spatiotemporal evidence that adds up to the neurodevelopmental hypothesis proposing different brain phenotypes, mainly differentiating CM from TM and CW.