Functional magnetic resonance imaging (fMRI) methods have been extensively applied to study the human brain and its functional organization in healthy and disease states. A strong rationale exists for the extension of this approach to spatially-resolve large-scale neuronal circuits in animal models and assesses the way these mediate specific behaviours. The combination of structural and functional MRI defines a novel powerful translational paradigm to bridge clinical and preclinical research of brain pathology.
Specific projects we are currently pursuing include:
Mapping the mouse brain connectome
We recently provided the first description of the presence of robust, distributed resting-state networks in the mouse brain using fMRI (Sforazzini et al., submitted). This research lines aims to use high resolution diffusion tensor imaging (DTI) and resting-state fMRI to obtain a description of the intrinsic connectional architecture of the mouse brain. The overall goal is to be able to obtain a large-scale, integrated portrayal of the mouse connectome, i.e. the complex network of elements and connections that govern and compose the mouse brain. This project parallels analogous research efforts in humans, and will be ultimately aimed to determine the elusive relationship between structural and functional connectivity, as well as to identify large-scale descriptors of disease-related derangements in brain connectivity, a feature that appears to be prominent in brain disorders such as autism and schizophrenia.
Resting-state fMRI in the mouse (a) revealed distributed antero-posterior networks encompassing prefrontal, parietal and peri-hippocampal cortex. The approach can be used to describe the connectional architecture of the brain (b; cross-correlation matrix between cortical and sub-cortical areas; c; graph representation of the matrix in b). From Sforazzini et al., submitted.
Linking circuits to behaviour: optogenetic and pharmacogenetic modulation of brain activity
The research line aims to use optogenetic and/or pharmacogenetic tools in combination with fMRI to dissect the circuital basis of behaviour. When combined with measurements performed in freely-moving animals, the technique has the potential to map the downstream circuits engaged by specific neuronal population and unambiguously associate this signature to specific behavioral or pathological traits. An illustrative example of the approach has been recently published by the lab (Gozzi A. et al., Neuron, 2010; 67: 656-66)
Pharmacogenetic silencing and fMRI were used to identify a novel circuit that controls the quality of fear responses. Transient silencing of the central nucleus of the amygdala was found to activate basal forebrain cholinergic nuclei (a) which in turn recruit cortical areas (b). The effect was associated to a switch between passive to active fear response (Gozzi et al., Neuron, 2010).
Mapping altered neurofunctional states with MRI
The laboratory has developed a rich repertoire of morpho-anatomical (e.g. VBM, cortical thickness mapping, DTI and white matter tractography) and functional MRI readouts (resting-state fMRI, pharmacologically-evoked fMRI) to provide a fine-grain multidimensional description of altered neuro-functional states in mouse models of brain disorders, with a particular focus on connectopathies (e.g. schizophrenia and autism). The use of readouts commonly used in the clinic offers a means to directly assess analogous alterations in clinical populations, and to objectively determine the translational relevance of the preclinical models.
Major white matter reorganization was observed in BTBR mice (a), an inbred mouse strain used to mimic symptoms of autism, using high-resolution diffusion tensor tractography. The effect was associated to a generalized reduction in cortical thickness (b). From Dodero et al., submitted.
The laboratory has a successful record in the application of pharmacological fMRI (phMRI) to describe the central substrates recruited by neuro-active agents belonging to different psychopharmacological classes and, more recently, to map the central substrate of endogenous d neuro-modulators and intranasally-administered neuro-peptides. The use of drug-elicited fMRI responses to assess the mechanistic efficacy of novel pharmacological agents is an emerging clinical paradigm that has been pioneered in our lab (Bifone A & Gozzi A, Expert Opin. Drug Discov. 2012).
Pattern of activation produced by a clinically-relevant dose of the pro-cognitive drug Modafinil in the rat brain. The recruitment of regions involved in cognition is apparent (from Gozzi et al., Neuropsychopharmacology, 2011)
Unsupervised classification of Multidimensional MR datasets
Image processing and analysis of multidimensional fMRI and DTI datasets pose formidable computational challenges in terms of data reduction and extraction of relevant neuroanatomical and functional information. This is particularly true when these datasets are employed to make inferences between control and diseased populations. Our group is collaborating with external groups of established reputation graph-based and machine learning approaches (Prof Murino, Pavis, IIT Genova, Prof Guido F. Caldarelli, IMT, Lucca) in an attempt to define unsupervised classification schemes for multidimensional datasets.
Automatic identification of major white matter bundles in the mouse brain using dominant sets (Dodero et al., PRIN 2013)
The goal of this activity is to simultaneously record anatomical and functional MRI signals in vivo with complementary investigational techniques, such as multi-electrode electrophysiology, optical methods and metabolic probes (tissue oxymetry) to validate MRI signals of interests. Hybrid optical and MR systems may allow the combination of molecular targeting capacity imparted by fluorescent reporter probes with the versatile tissue-function contrast achieved with MRI. Similarly, the combination of electrophysiological or tissue oxymetry signals with fMRI measurements can greatly help elucidate the elusive origin of positive and negative haemodynamic responses.
The lab also seeks to apply the technology to other application domains within the IIT network and through external collaborations with academic third parties.