Researchers at the Champalimaud Foundation have taken a significant step towards unravelling how information travels through the brain. Using a new Ultrafast MRI technique (UF-MRI), they were able to determine for the first time the direction in which nerve impulses travel: from the senses to higher brain areas (feedforward) or, conversely, from expectations and predictions (feedback).
The study was conducted on rat brains and offers new insights into fundamental brain processes. This discovery could have significant implications for understanding neurological and psychiatric conditions such as autism, Alzheimer’s, schizophrenia and hallucinations. According to lead researcher Noam Shemesh, this represents a new way of looking at brain activity: not just where activity occurs, but how information actually flows.
Making invisible processes visible
Central to the research is an innovative imaging technique: uFLARE (UltraFast Layer-Resolved Encoding). This method makes it possible to measure brain activity with an unprecedented level of temporal and spatial resolution.
The researchers used an experimental MRI scanner with a magnetic field of 9.4 Tesla, considerably more powerful than the 1 to 3 Tesla typically used in clinical settings. This enabled them to produce functional MRI images with a resolution of milliseconds and at the sub-millimetre level. The results of the research have been published in Nature Communications.
This speed and precision are crucial for distinguishing between different directions of signal transmission in the brain. Even spontaneous activity, which occurs continuously, even at rest, was found to exhibit clear, distinctive patterns for feedforward and feedback signals. However, imaging alone was not sufficient. To interpret the complex datasets, the researchers developed additional computational models capable of reconstructing the direction of information flows.
The cerebral cortex as a layered network
The analysis focused on the visual cortex, a brain region consisting of six clearly distinct layers. This structure makes it possible to accurately determine where different types of signals enter and exit.
Feedforward signals, originating from the senses, are directed primarily towards the middle layer of the cortex. Feedback signals, originating from higher brain areas, project instead towards the superficial and deeper layers. This anatomical knowledge formed the basis for the model developed.
The team used a so-called “layer-based connective field model” to describe communication between neurons. This model predicts how activity in one brain region influences another region, depending on the layer in which the signals are processed. Researcher Joana Carvalho explains that the size of the so-called ‘connective field’ indicates how much information is exchanged between neuronal groups. By analysing this variation across the different layers, the researchers were able to distinguish between ascending and descending signals.
The results confirmed the hypothesis: the patterns in the MRI data matched the model’s predictions. This enabled the researchers to reliably determine the direction in which information flowed through the brain.
Visual cortex
Although the research focused on the visual cortex, the method also proved applicable to other brain regions, such as the somatosensory and motor systems. This points to a possible general principle of brain communication. The implications for medicine are significant. Many neurological disorders are associated with disruptions in the balance between feedforward and feedback processes. Until now, however, it has not been possible to measure these changes directly.
According to Shemesh, the new method offers an opportunity to understand this better. “We know that communication pathways in the brain change in disorders such as Alzheimer’s or Parkinson’s, but we do not know exactly how. With this technique, we can now map those changes.”
In autism, for example, it is assumed that differences in perception are linked to abnormalities in how information is integrated. The new technique can help determine whether these differences stem from changes in feedforward or feedback signals.
Practical application
The next step for the research team is to translate these findings into studies involving humans. If the method proves applicable to the human brain as well, this could lead to new diagnostic possibilities and treatment strategies. According to Carvalho, this is where the real potential lies: “If we can see how lesions or disorders influence the direction of brain signals, we gain a much better understanding of the underlying mechanisms.”
This development highlights the growing role of advanced imaging, data analysis and computational models in neuroscience. Combining these technologies is yielding a new generation of insights into brain function.
The ability not only to measure brain activity but also to determine the direction of information processing marks a significant step towards precision medicine in neurology. It opens the door to a future in which brain disorders can be better understood, detected earlier and treated more effectively.
AI tool improves brain scans
Last year, researchers at the Mayo Clinic developed an AI tool, StateViewer, which can significantly improve the diagnosis of dementia. Based on a single FDG-PET brain scan, the tool can distinguish between nine different forms of dementia. In studies, StateViewer identified the correct diagnosis in 88 per cent of cases and worked faster and more accurately than traditional methods.
The AI analyses glucose metabolism in the brain and compares scans with an extensive reference database. The results are displayed in intuitive, colour-coded brain maps, enabling even non-specialists to interpret the findings more effectively. This development could contribute to more personalised diagnoses, which is crucial given the growing number of people with dementia.
