2025 Topic-Specific Grants and Descriptions

To navigate the dynamic world, our brain needs to continuously update its representation of external information and generate predictions of future states. Without such predictions, there would be a substantial time lag between states in the real world and the perception of, and reaction to, these states. A fundamental assumption is that the brain constantly generates and updates internal models of the world. However, the representational nature of internal models at different processing levels, and how the dynamics of internal models temporally relate to (e.g. follow or predict) actual events in the real world, remains unknown. 

The objective of this PhD project is to investigate the representational dynamics in the brain in response to dynamic events (e.g. observed or executed actions) using a novel, magnetoencephalography (MEG)-based approach – dynamic representational similarity analysis (dRSA hereafter). The innovation of the approach is to use temporally variable models of representational similarity to characterize representational content at each time point during temporally extended, unfolding events. This allows testing whether a given time point is represented in a lagged bottom-up manner or in a predictive top-down manner, that is, before it actually occurred (for reference to the method, see https://www.nature.com/articles/s41467-023-39355-y). 

The successful candidate should have a background in cognitive neuroscience and neuroimaging (preferably M/EEG or fMRI, MVPA/RSA) and strong programming skills (preferably Matlab). Candidates are strongly encouraged to get in contact with Moritz Wurm.

Declarative knowledge, meaning the portion of knowledge that we can consciously access and manipulate, is one of the most enduring mysteries of the human mind. How did it evolve? And what are the mechanisms behind it? One possibility is that the complex neural machinery that mammals evolved to navigate space has been recycled to “navigate” declarative knowledge. Research from single-cell recordings in rodents to brain imaging studies with humans is converging toward the fascinating hypothesis that conscious declarative knowledge is spatially organized, and can be stored, retrieved and manipulated through the same computations used to represent and navigate physical space. 
In the last few years my lab has been testing and developing this hypothesis, relying upon cutting-edge neuroimaging and analysis techniques. We provided evidence that conceptual knowledge is organized within cognitive maps across complementary allocentric and egocentric reference frames; We studied the role of vision in the development of cognitive maps, showing a different neural geometry underlying spatial navigation in Early Blind people; We provided evidence that abstract cognitive maps are navigated through internal attentional movements; among other contributions. 
In this new project, building upon our recent findings, we aim to decode the neural mechanisms underlying relational declarative memory, working at the intersection of attention, memory, and consciousness studies. Our investigations will utilize intracranial EEG, involving electrodes temporarily implanted in the brains of patients with epilepsy. This approach offers the exceptional advantage of achieving high temporal and spatial resolution, allowing a detailed characterization of the neurophysiological mechanisms supporting memory and high-level cognition. Due to the necessity of effective communication with patients and medical personnel, fluent knowledge of Italian is particularly important.

See below some of the recent publications from the lab (www.bottinilab.com):
Sigismondi, F., Xu, Y., Silvestri, M., & Bottini, R. (2024). Altered grid-like coding in early blind people Nature Communications
Dutriaux, L., Xu, Y., Sartorato, N., Lhuillier, S. & Bottini, R. (2024). Disentangling reference frames in the neural compass Imaging Neuroscience
Viganò, S., Bayramova, R., Doeller, C., & Bottini, R. (2024). Spontaneous eye movements reflect the representational geometries of conceptual spaces PNAS
Viganò, S., Bayramova, R., Doeller, C., & Bottini, R. (2023). Mental search of concepts is supported by egocentric vector representations and restructured grid maps Nature Communications
Giari G., Vignali L., Xu Y. & Bottini R. (2023). MEG frequency tagging reveals a grid-like code during attentional movements Cell Reports
Xu,Y., Vignali, L., Sigismondi, F., Crepaldi, D., Bottini, R., Collignon, O. (2023). Similar object shape representation encoded in the inferolateral occipitotemporal cortex of sighted and early blind people Plos Biology
Bottini, R., Doeller, FC. (2020). Knowledge across reference frames: Cognitive maps and image spaces Trends in Cognitive Sciences

Supervisor: Roberto Bottini

The ability to imitate has long been considered a unique trait of primates. Though increasing evidence from various animal species challenges this idea, there are still very few studies that demonstrate this ability at birth, and even fewer that show it occurring without any prior experience. The project aims to investigate the neural substrates of spontaneous social imitation at birth prior to any visual experience using the domestic chick (Gallus gallus domesticus) as developmental animal model. To this aim, we will take advantage of classical behavioural paradigms such as the passive-avoidance learning paradigm, combined with neurobiological techniques - qPCR, immunohistochemistry, histology (e.g.: mitochondrial staining NBTx, Black Gold staining for myelin) and in-situ hybridization - and with neuroimaging techniques - MEMRI, DTI, non-anesthetized fMRI.
The ideal candidate should possess a Master degree in Psychology, Biology, Neuroscience or related disciplines and experience with behavioural and neurobiological methods as applied to birds (especially, to domestic chicks).

Supervisor: Giorgio Vallortigara

The economic impact of psychiatric disorders is projected to surpass that of cancer, diabetes, and respiratory diseases combined, highlighting a significant health crisis that demands urgent action. The advent of Magnetic Resonance Imaging (MRI) and functional MRI (fMRI) enabled the non-invasive measurement of structural and functional principles of brain organization across both health and disease, without however achieving clinical translation. The hypothesis at the core of the PhD project is that the universally adopted pre-processing strategies and analytical approaches are hampering the development of such translational tools. 

The primary aim of the project [AIM 1] is to develop an artificial intelligence-based solution that bypasses canonical preprocessing strategies and analytical approaches. If successful, the PhD applicant will provide a radical new solution to model inter-individual variability, overcoming the major pitfalls of the state-of-the art.

Akin to the revolution sparked by the introduction of general-purpose artificial intelligence models – also known as foundation models – the project aims to develop a similar framework to model interindividual variability in brain connectivity. The proposed architecture will capture structural, morphological and functional aspects of the brain’s intrinsic organization, analogously to the ability of foundation models to process audio, text, and images. 

In keeping with the possibility of foundation models to be easily adapted to the users’ needs, the secondary aim [AIM 2] of the project is to investigate model fine tuning in psychiatric settings. The research questions are devoted to probe the overall clinical validity, the investigation of ethnic and gender biases, and the feasibility of predicting the long-term trajectory of mental health conditions.

The PhD scholarship will be hosted at The Neuroinformatics Laboratory (https://nilab.fbk.eu/research/), directed by Dr. Paolo Avesani. The lab is embedded within the Center for Augmented Intelligence at the Fondazione Bruno Kessler (Trento). The supervisor will be Dr. Ludovico Coletta. 

The successful candidate is expected to develop from scratch point cloud and graph neural networks using python based solutions (PyTorch). Knowledge in neuroimaging – and especially of preprocessing softwares using the command line (FSL, AFNI, ANTs, and the Python ecosystem) – is a strong plus. For further information, candidates are strongly encouraged to get in contact with Paolo Avesani and Ludovico Coletta. 

The Laboratory of Autism and Neurodevelopmental Disorders at IIT (IIT-LAND), directed by Dr. Michael Lombardo, at the Center for Neuroscience and Cognitive Systems, Istituto Italiano di Tecnologia, Rovereto, invites applications for 2 PhD scholarships to investigate early biomarkers behind autism subtypes. Our research aims to understand how autism may be split into distinctive types of autisms, characterized by distinctive phenotypic presentation, underlying neurobiological mechanisms, and differential responses to treatment. To answer these types of questions, we use a combination of approaches from neuroimaging (EEG, MRI), cognitive and computational neuroscience, and data science. Our work is primarily focused on human patients (i.e. autistic children), but we also collaborate with other groups on larger cross-cutting translational work focused on elucidating biological mechanisms in model systems. For more info, see:  https://land.iit.it.

We have 2 PhD positions focused on identifying biomarkers for different clinically and behaviorally distinctive subtypes of autism. The work heavily focuses on methodologies such as eye tracking, EEG, and MRI/fMRI. The datasets we work with are a combination of large publicly available datasets, datasets from international collaborators, as well as data coming from ongoing experiments run within IIT-LAND. The work will heavily rely on new or past stratification models of neural and phenotypically distinctive autism subtypes. The two positions are fully-funded off of an ERC Consolidator Grant to Dr. Michael Lombardo.

We are looking for talented and highly motivated individuals that can build on prior skill sets or knowledge within the core areas of our research - EEG, eye tracking, neuroimaging (MRI, fMRI), data science, and autism. Advanced understanding and conceptual thinking with regards to statistics and big data analysis is a plus, as is requisite computational and programming skills (e.g., R, Python, MATLAB) to implement such ideas. Ability to speak both Italian and English is highly prioritized. Good communication skills and ability to work within larger groups is also emphasized. Strong passion/desire to pursue an academic career focusing on neurodevelopmental disorders like autism is also a key characteristic we are looking for, as is an existing strong grasp of the literature on autism and neurodevelopmental disorders.

Successful candidates will join a growing diverse, multidisciplinary and collegial group, and will be offered direct training, supervision and mentorship from the principal investigator and other senior members of the lab. The work also offers up the possibility of working within a larger international context, as nearly all of the work we do on this topic is done with collaborations from colleagues in Europe and the USA. The position is a four-year scholarship in the international doctoral school in Cognitive and Brain Sciences (CIMEC) at the University of Trento. Candidates will join a diverse cohort of PhD students and receive multi-disciplinary training at the interface of computational, experimental and cognitive neuroscience across humans and animal models.

The IIT-CNCS in Rovereto is actively expanding its infrastructure for systems level neuroscience research. Our center is located in Trentino, a region of Northern Italy nested within the Dolomite mountains, offering easy access to spectacular natural beauty and mountaineering, vibrant culture and exceptional quality of life (https://www.iit.it/it-IT/cncs-unitn/).

Candidates can informally contact Dr Michael Lombardo (michael.lombardo@iit.it) to gather more information about the position, the project, the application procedure and the selection process.

This project aims to reveal the unifying neural principles supporting visual perception across species, by linking psychophysics in humans and mice with large scale neural recordings. It is supported by a collaboration between the labs of Peter Neri and Federico Rossi, both at IIT. Peter Neri (sites.google.com/site/neripeter) is an expert in human psychophysics and computational neuroscience, and leads the Sensory Processing and Neural Computation Lab at IIT Genoa. Federico Rossi (rossilab.iit.it/people) is an expert in visual physiology and large scale recordings in mice, and leads the Functional Architecture of Neural Circuits Lab at IIT Rovereto. Their combined expertise spans human psychophysics, computational modeling, and circuit physiology.

This project will reveal the neural geometry supporting visual perceptual judgements by linking  large scale multi-area population recordings with quantitative measurements of sensory behavior within a fully specified perceptual geometrical framework (Figure 1). The connection between human visual perception and the underlying neural circuitry remains elusive. A key challenge is bridging human studies with experimental neurophysiology to test if the representations predicted by behavioral modeling actually exist in the brain. Based on extensive behavioral data in humans, we have developed a framework for translating visual perceptual judgments into distance measurements on a surface representation in neural space. To determine whether the geometry inferred from behavior is represented at the level of neural activity, we will record from neuronal populations across visual areas in the mouse cortex with large scale 2P imaging while animals are engaged in sensory tasks that are virtually identical to those used in humans. Our approach will bridge computational psychophysics and experimental neurophysiology to develop models that can explain animal perceptual decisions. We are looking for candidates with an MSc in neuroscience, physics, engineering, or any STEM discipline. Given the interdisciplinary and technical nature of this project, you should be highly motivated to learn new methodologies in different areas spanning behavioral quantification, computational modeling, and experimental neurophysiology. Furthermore, we value creative self-driven individuals, so we would encourage you to develop your own extensions of the project outlined above.

References

Neri P (2018) The empirical characteristics of human pattern vision defy theoretically-driven expectations. PLoS Computational Biology 14(12): e1006585. Neri P (2024) Local geometry of elementary visual computations. MODVIS 2024 Abstract (Computational Models of Edges & Contours). https://docs.lib.purdue.edu/modvis/2024/cmec/2/ SueYeon C, Abbott LF (2021) Neural population geometry: An approach for understanding biological and artificial neural networks. Current Opinion in Neurobiology 70: 137–144. Rossi, L.F. , Harris, K.D., Carandini M. Spatial connectivity matches direction selectivity in visual cortex, Nature, (2020).

For more details, please see the lab webpage: https://www.rossilab.iit.it

Our lab aims to understand how the specialised activity of individual neurons, and brain areas, emerges from the architecture of their synaptic connections; how gene expression predisposes the blueprint of these networks; and how their plasticity allows animals to learn. To address these challenges, we investigate the neural architectures that enable adaptive visually guided behaviours in the mouse, leveraging methods from system neuroscience, physiology, anatomy, and molecular biology to link function, connectivity, and gene expression in vivo at multiple scales. For more info, see: https://rossilab.iit.it/

Neurons have mesmerizing dendritic trees, whose purpose has puzzled scientists for decades. Individual dendrites may operate as an independent processing unit of specialised synaptic inputs, capable of driving the soma with active properties, acting as coincidence detectors or gating drive from other dendrites. This functional compartmentalisation is thought to expand neuronal computational capabilities, and has inspired powerful models of neuromorphic computing. More importantly, dendrites are the substrate of the synaptic plasticity that allows neurons to learn new computations and form new memories. However, so far, it has been difficult to test these theories with causal, dendritic recordings and manipulations in vivo; therefore their role for information processing and learning remains unclear. 

To test these theories, this project will develop methods to simultaneously map and perturb synaptic inputs and output in single neurons, and deploy them to causally probe the function of dendrites in vivo during learning and adaptive visual behaviour. These include simultaneous synaptic glutamate and calcium imaging, optogenetics, optical pruning, and high- density electrophysiology. We will also guide and complement this experimental work with analysis and modelling of large scale functional connectomics datasets (https://www.microns-explorer.org/). 

To complete this project, we are looking for candidates with an MSc in neuroscience, molecular biology, physics, engineering, or any STEM discipline. Hands-on training in experimental neuroscience or molecular biology, and/or familiarity with programming (e.g. Python, Matlab), will be highly valued. Successful candidates will join a growing diverse, multidisciplinary and collegial group, will help build our new lab, and will be offered direct training, supervision and mentorship from the principal investigator.
For more details, please see the lab webpage: https://www.rossilab.iit.it

Selected References

Two journal issues (several papers):
https://www.sciencedirect.com/journal/neuroscience/vol/489
https://www.nature.com/collections/bdigiaicbd

Aggarwal, A., Negrean, et al., Glutamate indicators with increased sensitivity and tailored deactivation rates, bioRxiv (2025)

Ye Z, Shelton AM, et al., Ultra-high density electrodes improve detection, yield, and cell type identification in neuronal recordings. bioRxiv, 2024
 
Rossi, L. F., Harris, K. D. & Carandini, M. Spatial connectivity matches directionselectivity in visual cortex. Nature, 2020
 
Pachitariu, M. et al. Suite2p beyond 10,000 neurons with standard two-photon microscopy. biorXiv, 2016.

Learning relationships among objects, categories, and events encountered in our environment is critical for survival. In the mammalian brain, structured knowledge is believed to emerge from cognitive maps encoded within the hippocampal formation and the prefrontal cortex, through the coordinated activity of specialized neurons such as goal progress cells, place cells, and grid cells. Despite significant progress, the exact computations performed by these neurons and their roles in the acquisition, storage, and retrieval of structured information remain poorly understood. This PhD project aims to clarify these processes by combining advanced methodologies to study the rodent brain at the single-cell and neural population levels. The candidate will: Utilize state-of-the-art recording techniques to monitor neuronal activity. Analyze and categorize spontaneous animal behavior during categorization and decision-making tasks. Implement and validate computational models based on reinforcement learning and predictive coding frameworks. This integrative approach will provide novel insights into the neural basis of structured knowledge and predictive behavior in mammals.

Supervisor: Giuliano Iurilli