top of page

Do We Have Food On Our Brain?

A greasy pizza slice. Crispy French fries in a heap. On a steamy summer day, ice cream dripped off a cone. MIT neuroscientists have discovered that when you look at any of these items, a specific area of your visual brain lights up.

Along with populations that respond particularly to faces, bodies, locations, and words, this newly identified community of food-responsive neurons is situated in the ventral visual stream. The researchers presume that the unanticipated outcome may be a reflection of the distinct region that food contains in human civilization.

The research, which was based on an examination of a sizable public database of brain activity in reaction to a collection of 10,000 photographs, raises many more inquiries about how and why this neuronal population evolves. The researchers intend to investigate how people's reactions to particular foods may vary based on their preferences and knowledge of particular food categories.

Meenakshi Khosla, an MIT postdoc, and N. Apurva Ratan Murty, an MIT research scientist, are the paper's primary authors which h was published in the journal Current Biology.

Kanwisher found cortical areas that respond specifically to faces more than 20 years ago while researching the ventral visual stream, the area of the brain that detects things. Later, she, as well as other researchers, found other areas with selective responses to objects, people, or speech. When researchers actively set out to seek them, they found most of those regions. However, Kanwisher alerts that using a hypothesis-driven technique might limit the consequences you discover.

There could be additional things out there that we haven't thought to check for, she says. And even if we do discover anything, how can we be certain that it is a genuine component of that pathway's fundamental dominating structure rather than merely a byproduct of our search?

Kanwisher and Khosla decided to examine a sizable, publicly accessible database of full-brain functional magnetic resonance imaging (fMRI) answers from eight human individuals while they viewed millions of pictures in an effort to elucidate the underlying organization of the ventral visual stream.

"We wanted to examine what kind of selectivities emerge when we use a data-driven, hypothesis-free technique and whether they are congruent with what has been observed previously.”

The decreased spatial resolution of the fMRI data made it challenging to discern previously postulated selectivities, thus that was our second objective, according to Khosla.

To accomplish this, the researchers used a mathematical technique that allowed them to determine neuronal residents from non-traditional fMRI data. Numerous voxels, which are three-dimensional components that represent a cube of neural tissue, make up an fMRI picture. The hundreds of thousands of neurons that make up individual voxels may be flooded out by separate groups in the exact voxel if some of those neurons are part of slighter populations that only respond to specific classes of depicted input.

The novel analytical technique can reveal responses of neuronal populations inside each voxel of fMRI data, as Kanwisher's team has done before with data from the temporal lobe.

The scientists discovered four populations using this method that matched previously found groups that react to faces, locations, bodies, and language. According to Kanwisher, "it shows us that this strategy works and that the things that we uncovered earlier are not simply obscure aspects of that pathway, but important, dominating properties."

It is interesting to note that a fifth group also arose, and this one seemed to like representations of food.

Food is not a visually uniform category, so this initially confused us, claims Khosla. We discovered a single group that reacts accordingly to all these various food products, despite the fact that things like apples, maize, and pasta all seem so dissimilar.

The ventral food component (VFC), which the researchers refer to as is made up of two clusters of neurons that are situated on each side of the FFA. According to the researchers, the fact that the nutrition-specific populations are dispersed among other category-specific communities may assist to explain why they have not been observed before.

Because the communities that are chosen for food mix with other neighboring populations that have different responses to other stimulus qualities, we believe that food selectivity had previously been more difficult to quantify. Because the responses of many brain populations mix in a voxel, the poor spatial resolution of fMRI preclude us from recognizing this selection, claims Khosla.

Works Cited:

Anne Trafton | MIT News Office. “These Neurons Have Food on the Brain.” MIT News | Massachusetts Institute of Technology,

“The Center for Brains, Minds & Machines.” These Neurons Have Food on the Brain [MIT News] | The Center for Brains, Minds & Machines, 25 Aug. 2022,

“Neurons That Have Food on the Brain.” NSF,



bottom of page