smart people think faster than other people when solving problems? New findings by researchers at the Human Brain Project at Charité University Berlin and their partners at the University Pompeu Fabra in Barcelona call into question this deeply held view in the field of intelligence research.
Their research was recently published in the journal Nature Communications.
They constructed 650 personalized brain network models (BNM) from a biological point of view. The models, created using data collected from the Human Connectome Project, allowed the research team to simulate the processes the brain goes through during problem solving. Observations from the
brain simulations were compared with empirical data from 650 participants who took the Penn Matrix Reasoning Test (PMAT), which consists of a series of pattern-matching tasks of increasing difficulty. These results were quantified as participants' fluid intelligence (FI), which can be roughly described as the ability to make difficult decisions in novel situations.
"We found that people with high fluid intelligence (FI) took more time to complete more difficult tasks than those with low fluid intelligence (FI). They were simply faster at answering easy questions," explained the study's senior author, Petra Ritter of the University of Charité. "We first observed this in simulation experiments before finding that empirical data from IQ test participants matched this trend." Ritter's lab and many other research groups at the HBP use brain simulations to supplement observational data in order to build a theoretical framework of how the brain works. "
" In this case, brain simulations were used to identify links between brain functional and structural connectivity and cognitive performance. A more synchronized brain is better at problem solving, but not necessarily faster. "With reduced synchrony, decision-making circuits in the brain reach conclusions more quickly, and higher synchrony between brain regions leads to better integration of evidence and stronger working memory," Ritter said. Intuitively this is not surprising: if you had more time and considered more evidence, you would invest more in solving the problem and come up with a better solution. Here, we not only demonstrate this empirically, but also show that the observed differences in performance are a consequence of the dynamics of individualized brain network models. We thus present new evidence that challenges a common notion of human intelligence. "
Previously established local circuit models of working memory (WM) and decision-making (DM), both important to intelligence, were plugged into the Virtual Brain (TVB), which provided a whole-brain-level simulation.
Simulations were run using a multiscale brain modeling approach; brain imaging data were processed using an automated container pipeline. Processing of highly sensitive brain data was performed in a secure virtual research environment on the EBRAINS Health Data Cloud. These techniques are made available to the global research community through EBRAINS. Instead of thinking how fast, it's about understanding how biological networks decide decisions to develop biologically inspired tools and robotic applications. Modeling brain dynamics for intelligent decision-making is therefore a promising approach to building intelligent applications. "We believe that biologically more realistic models may one day outperform classical artificial intelligence," Ritter said. "
-
-
-
-
-
-
-
-
-
-