IMPRS CoNI Summer School 2026

Understanding the organization and function of human brain connections can be enriched with an evolutionary lens. For example, we can narrow down putative functions by identifying cognitive capacities and associated brain structures which are or are not shared with other animals. The language pathways are well-established in neuroscience, encompassing what were considered human unique brain pathways. However, new research demonstrates that chimpanzees and humans, but not phylogenetically more distant monkeys, share key ‘language’ pathways. New behavioural research correspondingly demonstrates greater vocal complexity in chimpanzees than African monkeys. Only one natural communication system is considered combinatorially complex with respect to mapping complex structure to complex meaning. This is human language. With a limited sound set, we combine words into utterances, generating endless new and relational meanings. Most animals have a limited sound set that is largely fixed from birth, and per species produce few multi-signal utterances in which the meaning shifts compared to the composing signals. However, recent studies delving into vocal sequence production suggest a dramatically different pattern in chimpanzee vocal production. Chimpanzees demonstrate highly flexible abilities to combine calls with ordering and recombinatorial properties. Vocal combinations which show compositional-like structures, such that calls combined into utterances may disambiguate, add or generate new meanings compared to the composing calls. Like humans, ontogenetic development is protracted, with utterance length and diversity dramatically increasing until 10 years of age. Such a developmental trajectory and the population differences documented in sequence structure are both suggestive of social learning capacities. Taken together, our results place chimpanzee vocal capacities between those of humans and African monkeys, with implications for evolutionary changes to reorganization of human language tract homologues through the primate lineage.

The human brain exhibits systematic patterns of structural similarity between regions that reflect coordinated development. I propose these patterns form a "biosocial envelope" a framework integrating genetic programs with social-environmental influences to shape individual brain development. Within this envelope, developmental trajectories unfold along three genetically-established organizational axes: spatial proximity, sensorimotor-transmodal hierarchy, and dual-origin architecture. Social and environmental factors modulate development along these same axes, creating individual variation within shared constraints. This framework offers new insights into neuroplasticity and psychiatric vulnerability. Transdiagnostic analyses reveal that diverse neuropsychiatric disorders produce co-alterations along these shared organizational axes, with different conditions following distinct routes through common brain architecture. These patterns identify spatio-temporal risk points and suggest why certain brain systems are preferentially vulnerable across conditions. Critically, structural similarity patterns serve as active measurements of biosocial interdependencies, enabling prediction of individual disease risk and treatment response. This positions the biosocial envelope as a plastic, context-dependent scaffold linking brain structure to function across health and disease.

Questions on the neurobiology of language and its evolutionary precursors are some of the most challenging to address in the neural sciences. For instance, how does the human brain allow us to transform sensory signals into meaningfully structured symbols for communication via language? How has our understanding of the neurobiology of the human language system improved since the seminal discoveries by Paul Broca and Carl Wernicke? What do we understand of evolutionary conservation and divergence of neurobiological function related to human language in nonhuman primates or other species? At this International Max Planck Research School, I aim to provide a primer on how our understanding of the neurobiology of this remarkable human ability continues to evolve, including with the recent notion of an expanded ‘core language network’ distinct from yet capable of interacting with systems for cognition, such as those supporting memory. Launching from this idea, I overview studies that provide some of the pieces to the puzzle on how language evolved. I will conclude by asking the research school attendees to project themselves into the future, imagining what key steps they might take on this challenging yet endlessly fascinating empirical endeavor. Acknowledgements: My deepest gratitude to Prof Angela Friederici and colleagues at the Max Planck Institute for Human Cognitive and Brain Sciences for the years of empirical inspiration and fascination with the origins of a uniquely human capacity.

The primate brain is a visuomotor brain. Many of its organizational themes are thought to be to the adaptations of early primates to the niche of the small branches of trees. In this talk, I will illustrate how we use comparative neuroimaging to study the organisation of the primate brain, demonstrating both the general principles and the variations that occur in different branches of the primate family tree. Using newly developed techniques to quantify diversity in brain organisation across species, we can show how the temporal lobe expanded and become more complex in anthropoid primates (i.e., monkeys and apes) and became especially more connected in the human lineage.

In this presentation, I will investigate the evolution of the brain basis for the core language system (syntax) by comparing the human network with equivalent areas and connections in other primate species, especially baboons and chimpanzees. The results suggest a more gradual evolution than the prevailing theory suggests. Next, to verify that this evolution is truly continuous and not convergent, I will investigate the behavioural functions that these brain structures have evolved support. Here, the results suggest that this neural network evolved for social communication first, before subserving syntax in human evolution. As an outlook, ontogeny may mirror phylogeny, with prelinguistic children already possessing most of this network, possibly using it for social communication. This paves the way for later linguistic development as the brain network matures.

Understanding recovery from post-stroke language impairment requires moving beyond classical, region-based models toward a network-based framework that captures the dynamic reorganization of large-scale brain systems. Emerging evidence suggests that recovery is not solely supported by language-specific regions, but also critically depends on the integrity and adaptability of domain-general control networks, including the frontoparietal control network, dorsal attention network, and default mode network, that underpin flexible cognition and goal-directed behaviour. Developing mechanistic models of recovery therefore necessitates mapping how these domain-general systems interact with specialized language circuits to re-establish communicative function. In this talk I will highlight the brain network–based models of recovery that characterize recovery as a distributed, system-level process. By leveraging scalable, online language and cognitive testing platforms, we can now longitudinally assess multiple dimensions of cognition and communication at population scale. Paired with neuroimaging, these digital tools allow for comprehensive mapping of whole-brain functional networks and their contribution to language recovery trajectories. Together, these approaches promise to redefine our understanding of post-stroke plasticity, enabling the development of more predictive, personalized models of recovery that bridge fundamental neuroscience and clinical translation.

Language impairments following brain damage are often explained in terms of focal lesions, yet across conditions such as stroke, brain tumours, and neurodegenerative disease, symptoms frequently reflect disruption of distributed brain networks. In this talk, I will present a network-based perspective on language disorders, focusing on how damage to white matter connectivity reshapes the language network and contributes to variability in impairment and recovery. Drawing on evidence from diffusion MRI, tractography, and neuropsychological studies across clinical populations, I will show how moving beyond lesion location towards connectivity mapping provides a unifying framework for understanding language deficits, compensation, and disease progression.

Non-invasive brain stimulation (NIBS) offers powerful means to modulate neural activity and to enhance motor and cognitive functions safely and effectively. In this talk, I will provide an overview of current neuromodulatory approaches inspired by the brain’s orchestrated mode of operation, highlighting their potential to facilitate both motor and cognitive processes in healthy and clinical populations (for review Rektorova et al., 2025). Key techniques such as transcranial magnetic stimulation (TMS) and transcranial electrical stimulation (tES) will be introduced, with particular emphasis on recent developments in non-invasive deep brain stimulation (nDBS). Among these, transcranial temporal interference stimulation (tTIS) has emerged as a promising approach to reach and modulate deep brain structures non-invasively (e.g., Wessel et al., 2023; Beanato et al., 2024; Vassiliadis et al., 2024, 2025). I will further discuss recent technological and methodological advances enabling orchestrated neuromodulation—the coordinated stimulation of distributed brain networks—to promote functional recovery and cognitive enhancement (e.g., Raffin et al., 2025; Wessel et al., 2024). The presentation will introduce the underlying principles and mechanisms of these approaches, outline their translational potential for rehabilitation, and discuss remaining challenges and future directions toward personalized, clinically applicable interventions. <

References

Beanato E*, Moon HJ*, Windel F, Vassiliadis P, Wessel MJ, Popa T, Pauline M, Neufeld E, De Falco E, Gauthier B, Steiner M, Blanke O& , Hummel FC& (2024). Noninvasive modulation of the hippocampal-entorhinal complex during spatial navigation in humans Science Advances Nov;10(44):eado4103. doi: 10.1126/sciadv.ado4103.
Raffin E, Bevilacqua M, Windel F, Menoud P, Salamanca-Giron RF, Feroldi S, Zandvliet SB, Ramdass N, Draaisma L, Vuilleumier P, Guggisberg AG, Bonvin C, Fleury L, Huxlin KR, Beanato E, Hummel FC (2025). Boosting Hemianopia Recovery: The Power of Interareal Cross-Frequency Brain Stimulation Brain in press
Rektorova I, Puplikova M, Fleury L, Simko P, Brabenec L, Hummel FC (2025). Non-invasive brain stimulation for treating cognitive impairment in neurodegenerative and brain lesion disorders Nature Reviews Neurology Sep 16. Online ahead of print doi: doi.org/10.1038/s41582-025-01137-z
Vassiliadis P, Beanato E, Wessel MJ, Hummel FC (2025). Temporal interference stimulation for deep brain neuromodulation in humans Nature Biomedical Engineering in press
Vassiliadis P, Beanato E, Popa T, Windel F, Morishita T, Neufeld E, Duque J, Derosiere G, Wessel MJ, Hummel FC (2024). Non-invasive stimulation of the human striatum disrupts reinforcement learning of motor skills Nature Human Behavior Aug;8(8):1581-1598. doi: 10.1038/s41562-024-01901-z.
Wessel MJ, Draaisma LR, Durand-Ruel M, Maceira-Elvira P, Moyne M, Turlan JL, Mühl A, Chauvigné L, Park CH, Koch PJ, Morishita T, Guggisberg A, Hummel FC (2024). Multi-focal stimulation of the cortico-cerebellar loop during the acquisition of a novel hand motor skill in chronic stroke survivors The Cerebellum Apr;23(2):341-354. doi: 10.1007/s12311-023-01526-4.
Wessel MJ*, Beanato E*, Popa T, Windel F, Vassiliadis P, Menoud P, Beliaeva V, Violante I, Abderrahmane H, Dzialecka P, Park CH, Maceira-Elvira P, Morishita T, Cassara A, Steiner M, Grossman N, Neufeld E, Hummel FC (2023). Noninvasive theta burst stimulation of the human striatum enhances striatal activity and motor skill learning Nature Neuroscience Nov;26(11):2005-2016. doi: 10.1038/s41593-023-01457-7

Memory is a complex process involving the encoding, storage, and retrieval of information that guides behaviour and decision-making. Central to this process is the hippocampus, which supports memory through both local activity and interactions with distributed brain networks. However, studying the human hippocampus directly has long posed significant challenges due to its deep location and the limitations of non-invasive tools. Temporal interference (TI) stimulation offers a promising solution by enabling focal, non-invasive neuromodulation of deep brain structures. In this talk, I will first explain the principles of TI and how it can selectively target deep brain regions like the hippocampus. I will then discuss our results manipulating memory encoding during wakefulness and consolidation during sleep by modulating hippocampal activity and hippocampal-cortical communication. This work offers new opportunities to understand the mechanisms underlying memory formation and stabilisation.

This talk will present recent evidence demonstrating that Transcranial Ultrasound Stimulation (TUS) can induce transient, targeted neuroplastic changes in the human brain, with implications for both mechanistic research and therapeutic development. I will highlight TUS's precision in targeting deep brain structures, its value in probing neural circuits, and its emerging clinical potential in psychiatric disorders. Drawing on recent human studies; including investigations into decision-making and a novel clinical work in obsessive-compulsive disorder (OCD); I will present evidence of TUS's effects on task-related neural changes, behaviour and symptoms demonstrating its promise as a transformative neuromodulation tool. I will also talk about the biosafety of TUS.

This talk will start with the physical effects of ultrasound and how ultrasound interacts with mechanical sensitive ion channels in the brain, including the concept of strain. We will then discuss how ultrasound parameters might affect these interactions. We will then discuss how to model and/or measure these effects in humans. We will close with a discussion clarifying the difference between exposure and dose, as well as laying out the key safety concepts.

The NexGen 7 Tesla scanner at University of California, Berkeley (also known as the MAGNETOM Terra Impulse) was developed to reach ultra-high resolution fMRI of human brain by designing and implementing several advances in hardware at Berkeley in collaboration with Siemens Healthineers and MR CoilTech, LTD. This includes a head-only “Impulse” asymmetric gradient coil (200 mT m−1, 900 T m−1s−1) with an additional third layer of windings, a 128-channel receiver system which allows development of 64- and 96-channel receiver coil arrays to boost signal in the cerebral cortex while reducing g-factor noise to enable higher accelerations and a 16-channel transmit system to reduced power deposition and improved image uniformity. The scanner routinely performs fMRI studies at 0.4–0.85 mm isotropic spatial resolution to reveal cortical layer functional activity in 3D whole brain CBF VASO coverage. Currently, we are developing local insertable shim array technology to reach 4th order shimming in the limited bore space. Secondly, to reach ultimate intrinsic SNR we are developing dual layered receiver arrays with 72 and 112 loops incorporating transceiver technology which raises SNR in the middle of the brain. Historically, high-performance gradients required for high b-value diffusion imaging (b ≥ 3,000 s/mm²) have been available on 3T scanners, reinforcing the assumption that ultra-high-field (7T) MRI is less suited for advanced diffusion imaging despite its inherent SNR advantages. Our recent work challenges this paradigm, demonstrating that combining 7T ultra-high magnetic field strength and high-performance gradients results in a fundamental shift in achievable spatial and angular resolutions in human diffusion imaging. Leveraging multiplicative SNR gains from higher field strength, shorter diffusion gradient pulses and faster image encoding, has enabled the NexGen 7T scanner to achieve much higher SNR in ultra-high b-value diffusion imaging than the 3T Connectome scanners with 300 mT/m or 500 mT/m gradients. The incorporation of ultra-high-performance gradients into a 7T scanner enables new diffusion encoding regimes that simultaneously support sub-millimeter spatial resolution and high angular resolution, combinations that have not previously been achievable in vivo in human brain.

Concurrent TMS–fMRI was first introduced in 1998, demonstrating the feasibility of combining the two techniques at 1.5 T. This pioneering work generated widespread interest across psychiatry, cognitive neuroscience, neurology, and rehabilitation medicine, underscoring the potential of the approach to causally probe the effects of TMS on brain activity and brain function. Nearly three decades later, however, concurrent TMS–fMRI has yet to fully mature into a robust and broadly adopted scientific tool. Building on insights gained from both early and contemporary implementations, I will discuss the key remaining limitations and outline how dedicated, purpose-built hardware—such as the RF Cap—can help overcome these challenges and bring the methodology closer to realizing its full potential. In addition, I will present my ongoing work supported by the NIH BRAIN Initiative K99/R00 award, which aims to further advance concurrent TMS/fMRI technology through the integration of EEG. This RF-EEG Cap approach is designed to enable truly high-resolution causal brain imaging and to facilitate the translation of concurrent TMS/fMRI/EEG into more sophisticated clinical and neuroscientific applications.

Ultra-strong gradient diffusion MRI offers a powerful pathway toward resolving microstructural features of the human brain with unprecedented fidelity. The Connectome 2.0 platform, equipped with 500 mT/m gradients and advanced hardware engineering, enables access to shorter diffusion times, higher b-values, and more flexible encoding schemes than conventional clinical MRI systems. These capabilities substantially improve sensitivity to axonal geometry, soma size and density, and the complexity of fiber populations that underlie human brain connectivity. In this talk, I will highlight recent advances in acquisition design, hardware and software co-optimization, and model development that exploit the unique operating regime of Connectome 2.0. Applications include sub-millimeter-resolution diffusion tractography, enhanced estimation of neurite and soma density distributions, and quantitative assessment of microstructural changes relevant to aging and neurodegeneration. I will also discuss strategies for harmonizing the ultra-high-gradient imaging system with next-generation 3T platforms to facilitate translation to large cohorts and clinical studies. By extending diffusion MRI into a regime that markedly improves microstructural specificity and delineation of connectional anatomy in the living human brain, Connectome 2.0 provides a next-generation framework for mapping the architecture of the human brain at scales that directly inform systems neuroscience, network modeling, and comparative neuroanatomy.