We introduced standardised dMRI tractography protocols for delineating cortico-subcortical connections between cortex and the amygdala, caudate, putamen, and the hippocampus, across humans and macaques. Building upon our previous work (Mars et al., 2018b; Warrington et al., 2020), which already provided protocols for cortico-thalamic radiations, and guided by the chemical tracer literature in the macaque, we devised the new protocols first for the macaque and then extended to humans. We demonstrated that our reconstructed tracts preserve topographical organisation principles, as suggested by tracers (Haber et al., 2006; Haber, 2016; Oldham and Ball, 2023).

As outlined in Schilling et al., 2020, tractography reconstructions can be highly accurate if information about where pathways go, and where they do not go is available. This is the philosophy behind the proposed protocols, which provide this type of constraints across different bundles. At the same time, these constraints are relatively coarse so that they are species-generalisable. We found that the proposed approaches yield tractography reconstruction across a range of datasets and respect individual similarities stemming from twinship. We further assessed the efficacy of these protocols in performing connectivity-based identification of homologous cortical and subcortical areas across the two species (Mars et al., 2018b; Mars et al., 2021; Warrington et al., 2022).

Mapping WM tracts that link cortical areas with deep brain structures (subcortical nuclei and hippocampus), as done here, enhances capabilities for studying neuroanatomy in many contexts, from evolution and development to mental health and neuropathology. As one of the (evolutionarily) older brain structures, the subcortex modulates brain functions including basic emotions, motivation, and movement control, providing a foundation upon which the more complex cognitive abilities of the cortex could develop and evolve (Haber et al., 2006; Pennartz et al., 2009; Haber, 2016; Sherman, 2016; Cruz et al., 2023). This modulatory function is mediated via WM bundles (Haber, 2016; Chumin et al., 2022). Consequently, their disruption is linked to abnormal function and pathology, in mental health, neurodegenerative, and neurodevelopmental disorders (Heller, 2016; Peters et al., 2016; Weerasekera et al., 2024). For example, in depression, fronto-thalamic (Bhatia et al., 2018), cortico-amygdalar (Arnsten and Rubia, 2012; Jalbrzikowski et al., 2017), and cortico-striatal (van Velzen et al., 2020) connectivity changes have been reported, while in schizophrenia there are associated fronto-striatal (Levitt et al., 2017) and hippocampal connectivity (Ikeda et al., 2023) changes. In Parkinson’s, there is impairment in fronto-striatal connectivity (Theilmann et al., 2013; Von Der Heide et al., 2013; Khan et al., 2019; Marecek et al., 2024), while fronto-thalamic and cingulate connectivity are impaired in Alzheimer’s disease (Von Der Heide et al., 2013; Bubb et al., 2018). Connectivity between the frontal lobe and the amygdala, thalamus, and striatum, as well as cingulum connectivity, is impaired in obsessive compulsive disorder, autism spectrum disorder, and attention deficit hyperactivity disorder (Langen et al., 2012; Arnsten and Rubia, 2012; Haber and Behrens, 2014; Kilroy et al., 2022). Therefore, reconstructing connectivity of these deep brain structures (striatum, thalamus, amygdala, and hippocampus) in a standardised manner, as enabled by our proposed tools, allows for further investigation into a wide range of disorders.

In addition, tractography of connections linking to/from deep brain structures has been used or proposed for guiding neuromodulation interventions, for example, deep brain stimulation (DBS) (Haber et al., 2021; Alagapan et al., 2023) or repetitive transcranial magnetic stimulation (rTMS) (Peters et al., 2016). DBS can inherently target subcortical structures and connectivity of subcortical circuits can be used to identify efficacious stimulation targets (Pouratian et al., 2011; Akram et al., 2017). rTMS, on the other hand, modulates subcortical function indirectly by targeting the structurally connected cortical areas. For example, dmPFC has been targeted to modulate the reward circuitry, in cases of anhedonia, negative symptoms in schizophrenia and major depression disorder (Dunlop et al., 2020; Gan et al., 2021; Bodén et al., 2021), while the vmPFC has been used as a target to modulate the prefrontal–striatal network (part of the limbic system) and regulate emotional arousal/anxiety (Chen et al., 2020; Kroker et al., 2022; Moses et al., 2025). Our results show a good mapping across species of both these cortical regions with specificity in their connectional patterns. Additionally, the motor cortex has been used as a target to modulate cortico-striatal connectivity in general anxiety disorder (Balderston et al., 2020; Fitzsimmons et al., 2024). We thus anticipate that having a standardised set of tracts linking the striatum, the hippocampus, the amygdala and the thalamus (all potential sites for stimulation) to specific cortical areas can assist the planning of interventions.

Our cross-species approach naturally lends itself to the study of evolutionary diversity. A number of comparative studies have revealed differences and similarities when comparing brain connectivity between humans and non-human primates (Barrett et al., 2020), including macaques (Mars et al., 2018b; Warrington et al., 2022) and chimpanzees (Bryant et al., 2025). Our work naturally extends these efforts and provides new tools for studying this diversity in deeper structures and subcortical nuclei. The ever-increasing availability of comparative MRI data (Bryant et al., 2021; Tendler et al., 2022) allows the definition of similar protocols in more species, such as the gibbon (Bryant et al., 2020; Bryant et al., 2024) or the marmoset monkey, and even across geometrically diverse brains depicting different stages of neurodevelopment (e.g. neonates vs adults) enabling concurrent studies of phylogeny and ontogeny (Warrington et al., 2022).

Our protocols have been developed and tested using FSL-XTRACT, but, in principle, are not specific to FSL. We have not evaluated performance with other tools, but these standard-space protocols could be translated into other tractography approaches. As described before, the protocols are recipes with anatomical constraints, including regions to which the corresponding WM pathways connect and regions they do not, constructed with cross-species generalisability in mind. Caution may be needed, however, if applying such protocols for segmenting whole-brain tractograms, as these can induce more false positives than tractography reconstructions from smaller seed regions and may require stricter exclusions.

Despite the potential demonstrated in this work, our study has limitations. As this is the first endeavour of this scale to map cortico-subcortical connections in a standardised manner and across two species, it is not exhaustive. Tracts linking the cortex to the striatum were prioritised as they are of increased relevance in human development and disease. However, expanding to include more tracts targeting other structures would provide a more holistic view. Our protocols were developed in the adult human brain. Future work will translate them to the infant brain (expanding on previous work Warrington et al., 2022) to interrogate cortico-subcortical connectivity across development. Tractography validation is a challenge, as is validation for any indirect and non-invasive imaging approach. We explored and demonstrated the generalisability of the proposed protocols, both within and across species. We also showed how the imaging-based reconstructions follow topographical organisation principles suggested by tracers.