Investigating tiny subcortical sensory nuclei and their substructures in humans in vivo is technically extremely challenging. We have developed novel neuroimaging approaches which allow to investigate sensory thalami at an unprecedented level of spatial resolution in-vivo. For example, we were able to dissociate two major subsections of the visual thalamus (pLGN, mLGN) by their tissue properties in humans in-vivo using quantitative magnetic resonance imaging. This led to a publicly available LGN atlas that can now also be used by other research groups.
Using advanced neuroimaging methods, we showed how sensory thalami are involved in speech recognition. These findings imply that a full understanding of human speech recognition abilities needs to take dynamic corticothalamic interactions into account. In further experiments we revealed that sensory thalamus responses can be explained by predictive-coding – a leading theory of brain function. The results confirmed hypotheses derived from a novel model of cortico-subcortical interactions that was at the centre of the SENSOCOM project. We also developed a computational model that is in congruence with the experimental findings and explains how predictive coding is used for major acoustic components of speech.
Using the advanced neuroimaging techniques developed in SENSOCOM we were able to investigate long-standing, but never directly tested hypotheses in developmental dyslexia and autism. In dyslexia, we found specific alterations in the mLGN and a specific reduction of white matter tracts that connect left-hemispheric visual motion processing areas (V5/MT) to the LGN. We obtained analogous findings for the auditory modality, i.e. an reduction of white matter tracts that connect left-hemispheric motion processing areas (mPT) with the auditory sensory thalamus (MGB). The specificity of the alterations to these cortico-thalamic connections is an important finding for two reasons: First it is a clear indication that developmental dyslexia is characterised not only by cerebral cortex dysfunction as proposed in most current developmental dyslexia models. Second, it gives first insight into the potential underlying mechanisms of the cortico-subcortical alterations, because both in the auditory and the visual modality it is the structures involved in motion processing that are altered.
For autism, we found alterations in bilateral mLGN, but not pLGN, thereby confirming a longstanding hypothesis about alteration of this part of the visual pathway in autism. Unexpectedly, we also discovered, that speech and voice processing is associated with functional alterations in brainstem structures in autism. Together the results imply that several symptoms in autism, such as difficulties with processing communication signals can be atleast partly explained by alterations in sensory pathway structures. This finding is important as autism is often explained by complex cognitive and emotional alterations; our findings rather imply that one also needs to take potential sensory alterations into account.
The SENSOCOM project has led to several follow-up projects. One uses neurostimulation to investigate the influence of cerebral cortex areas on the sensory thalamus. The other aims at a mechanistic explanation of sensory thalamus alterations in dyslexia. Most notable, the results of SENSOCOM inspired an interdisciplinary cooperation between four labs with different model systems and approaches to investigate thalamus and developmental dyslexia (non-human primates, rodents, human adults and children). This cooperation received an ERANET-Neuron Grant (ReDyslexia). One of the objectives of ReDyslexia is to use results from the SENSOCOM projects to motivate a proof-of-principle treatment study for dyslexia. In addition, some of the projects of SENSOCOM are continued and will shed light on the composition of LGN-V5/MT tracts and the mechanistic explanation
The project results have been communicated in scientific publications, many press-releases, talks and posters at conferences, and public stakeholders.
