Visualizing the Brain

A NOVEL APPROACH

Published January 6, 2022 • 7 min read

The convoluted, wrinkly, and mysterious shape of the brain has boggled scientists and inspired wonderers for hundreds of years. While the advent of psychological and neurological studies have begun to demystify this gray matter, scientists have just barely skimmed the surface. If understanding the brain was analogous to a person walking a mile, they would have only taken a few steps in this exploratory journey.

The shape of the brain has been marvelously and meticulously crafted by evolution for millions of years. Increasing in size and complexity with the advent of new species and behaviors, the brain has been warped, expanded and compressed by the growing bodies that hold it. The human brain evolved in a sort of curve, as apes and hominids evolved to be upright. Spatial constraint within the skull inspired folds — sulci and gyri — in order to maximize surface area. Regions within the brain became specialized embedded circuits, performing in patterns relative to the rest of the brain, such that behaviors became complicated within social animals, even up to the point of human society. 

As neuroscientists sought for means to understand the brain, a leading theory spearheaded the field: behavioral function is regionally located within the brain. This began with the idea of phrenology. Phrenologists believed that superficial marks on the skull signified intellectual capabilities. While this is now understood as false, this theory inspired and directly led to the comprehension of regionally associated functions that we have today. 

Since the brain is still a frontier, much still needs to be researched, but the findings to date have been excitingly enlightening. VALUENEX, with its novel Radar visualization technology, has taken this research and yielded unrecognized knowledge from it, including insights that are invisible to a research paper surveyor. For the trial visualization, 10,000 PubMed research papers were pulled to produce a dataset of abstracts associated with subcortical brain structures. After running the data through the VALUENEX machine intelligence and visualization processes to gain a comprehensive layout of the documents based on semantic similarities, the Radar’s results are brain-boggling.

Fig • comparison of the radar analysis of ‘subcortical brain structure’ research papers (left) and a brain cross-section (right)

Utilizing the heat map capabilities of Radar to clearly define document density — with similar documents being grouped together and distance between groups representing the degree of semantic similarity —, we can begin to see alikeness and parallels between the general shapes. While both the cross-section of the brain and the Radar share semicircular shape, notice a central red area within the Radar that parallels the midbrain on the cross-section image. Also notice the sparse areas (regions of low density) on the Radar, which almost directly mirror the ventricles in the brain diagram. 

If the visual similarity is shocking, the keyword labels produced by the algorithm are electrifying. First, observe the area labeled with keywords “schizophrenia, thalamus”. This small area geographically corresponds directly to a region in the brain called the thalamus, a nuclear complex with multiple connections to different parts of the brain regions. Even more fascinating is the connection between this region and the disorder schizophrenia; research reveals that in schizophrenia patients, there exists decreased connectivity between the prefrontal cortex and the thalamus, as well as disruption in their thalamic resting state networks. The Radar was successfully able to identify subcortical regions of the brain and associate a chronic brain disorder that directly relates to the functional performance of this region.

Note how the region labeled with keywords “LGN, V1” corresponds to the locations of the lateral geniculate nucleus (LGN) and the primary visual cortex (V1) of the brain. This area is very near the true location of the LGN in a sagittal medial slice (a view of the brain sliced at the central fissure), yet slightly relocated towards the occipital cortex. The Radar was able to automatically generate the primary visual cortex title near the occipital lobe where it is actually located in the brain. The inclusion of the keyword ‘V1’ in this region suggests the relationship of these two regions. Neurons in the LGN provide visual inputs to the V1, which are then processed and passed onto other visual cortical areas. The location and combination of keywords within the Radar accurately communicate the functional geography of the brain itself. 

Furthermore, the sparse regions of the Radar are representative of the different ventricles of the brain. The sparsity of these regions within the Radar mirrors the fact that aside from cerebrospinal fluid, the brain ventricles are actually hollow. The ventricles are fluid-filled structures that serve to maintain the central nervous system and keep the brain buoyant. It is fascinating to see how the Radar recognizes the more passive function of the ventricles and assigns these regions to be sparse.

The similarities and crossovers between topics are so closely tied and deeply embedded via the functions of neighboring regions within the brain that the locations themselves are hardwired into the information. The brain’s shape is necessary and optimal for its function, especially considering that similar topics and behaviors are located next to each other and are heavily interconnected. It is almost as though regions of the brain are spheres within a voluminous, complex multi-Venn Diagram. 

The beauty of this visualization is not just in its profound mirroring, but in its use. The Radar can show both sparse areas and the evolution of research trends. This can reveal understudied topics, unexpected crossover-concepts, and reveals potential theses and research topics all while affirming other associations and theories. 

For example, some research on brain diseases and neural afflictions appears embedded in the Radar. Certain orbital regions are dedicated to cysts, HIV, or Alzheimer’s, which are issues not typically associated with subcortical brain structures. Research on depression and anxiety disorders has a surprising overlap with general amnesia and non-disease affiliated memory loss. This suggests a correlation, and therefore, this is a region for investigation. How do mental illnesses affect memory recall, and why might both of these regions be located just around the pituitary gland? 

Investigations into peculiarities such as these might lead to discoveries and real solutions for afflicted people. At the dawning age of neurotechnology, new crossovers on technology and the brain are imminent and world-changing. While people often expect such crossovers to stem from new devices such as Elon Musk’s Neuralink, changes in the industry are often unexpected. Coupled with the boom in information sciences, VALUENEX is at the forefront of a new neuroinfo industry. Novel data analysis such as the Radar provide an unexpected foundation for progress, and a surprisingly clear lens into the complex, intimidating, and foggy field of neuroscience.