Brain biometrics

The brain fingeprint (brainprint) / L' empreinte cérébrale

EPFL, 2021-10-18

Media articles:

(EN) "Our brains have a “fingerprint” too"

(FR) "Chaque cerveau possède sa propre empreinte"

Scientific article:

Van De Ville, D. et al (2021). When makes you unique: Temporality of the human brain fingerprint. Science Advances, 7(42). https://doi.org/10.1126/sciadv.abj0751

EU research service - EU CORDIS (Community Research and Development Information Service), 2023-10-21

"Just like fingerprints, we all have a brainprint"

Commentary:

According to the above study, in order to identify an individual optimally, it suffices to obtain a 200-second brain recording, have access to a brain database and conduct a statistical correlation test.

The "Brainprint", the "Fingerpring" of the brain: Patterns of Brain Connectivity Represent Biometrics

Our fingertips have distinctive patterns, called "fingerprints" which can be used to identify us. The fingerprint constitutes a biometric, meaning a biological measurement or a measurable biological characteristic that allows the identification of an individual. In recent years, high-quality measurements of an individual's brain have been obtained enabling the characterization of personality traits and behavior. Interestingly, researchers have suggested that patterns of correlation in brain activity between different regions may act as a kind of “brain fingerprint,” sometimes referred to as a "brainprint".

Neuroscience seeks to understand the structure and function of the brain, including what different brain regions do and how they interact with one another to support different aspects of cognition and behavior. To this end, researchers study correlations (or correlates), meaning relationships or connections between brain regions, which may be either structural or functional. Distant brain regions may be connected structurally, for instance via white matter tracts. They may also exhibit similar patterns of activity, indicating that they are connected functionally. The strength of this functional connection is referred to as functional connectivity (FC). The complete map of the structural and functional connections in the brain is known as the connectome. 

Functional connectivity (FC) describes the correlation between the (functional) activity of different brain regions. In other words, it reflects whether—and how strongly—the activity of different brain regions fluctuates together over time. When two regions fluctuate together, it suggests that they may be functionally linked and working together as part of the same brain network to support a particular brain function. To study this, scientists obtain measurements showing how the activity level or each brain region increases or decreases with time — that is brain activity time courses (series) for different regions, using fMRI. They then compute statistical correlations between these signals to determine how closely the activity of different regions is related. The strength of this relationship is quantified using a numerical measure called the correlation coefficient.

Scientists represent these relationships using a matrix—a table describing the correlations between all possible pairs of brain regions. The brain regions are listed both horizontally and vertically, and each cell contains the correlation coefficient, a number describing how strongly the activity of the two corresponding regions fluctuates together and is therefore correlated. This table is called a functional connectivity matrix (FC matrix). In essence, it provides a map of statistical relationships between brain regions. Because the overall pattern of correlations can be distinctive for each individual, these connectivity patterns can sometimes act like a “fingerprint” of the brain, sometimes referred to as a brainprint.

Studying these connectivity patterns using tools from network science is known in neuroscience as brain connectomics.

Brainprint studies

A major development in the field was the work of a Yale team [1] in 2015, within the framework of the Human Connectome Project (NIH Blueprint for Neuroscience Research). They were among the first to demonstrate that, to a large extent, it is possible to identify an individual's functional connectome from a functional connectivity (FC) database, simply by computing the connection-wise (Pearson) correlation between a target FC matrix and those in the database [2].

An EPFL team [2] in 2021, demonstrated that an optimal brain fingerprint can be obtained at a time scale of 200 seconds. However, individual "snapshots" are reported to emerge at much shorter time scales.

Which regions appear first and which appear later? The scientists discovered that subcortical regions are the fastest ones for individual identification, while visual and somatomotor regions come right after. Ultimately, at longer time scales, higher-order regions i.e. frontoparietal and default mode network (DMN) emerge. 

Both cited studies used data from the HCP database. As mentioned for the first study [*]: "Notably, no brain scans or psychological tests were required"; "instead, the researchers drew upon an unprecedented trove of shared data from more than a thousand subjects made available by the HCP."

[1] Finn, E. S. et al. (2015). Functional connectome fingerprinting: identifying individuals using patterns of brain connectivity. Nature Neuroscience, 18(11), 1664–1671. https://doi.org/10.1038/nn.4135

[2] Van De Ville, D. et al (2021). When makes you unique: Temporality of the human brain fingerprint. Science Advances, 7(42). https://doi.org/10.1126/sciadv.abj0751

"New Evidence Points to Personal Brain Signatures"

Media article:

"New Evidence Points to Personal Brain Signatures"

Scientific American, 2016-04-13

"Brain scans of a person doing nothing at all can predict how neural circuits will light up when that same individual is gambling or reading a book"

Scientific article:

"Task-free MRI predicts individual differences in brain activity during task performance"

The following prior study by Finn et al. is cited with the comment that brain networks contain enough information to identify individuals with 99% accuracy: "Brain scans pinpoint individuals from a crowd"

Brain folds resemble fingerprints

Troubles neurologiques : élucider les mystères des plis du cortex cérébral

CNRS, 2024-11-03