Regions in the human inferior temporal gyrus are engaged in numerosity processing across visual stimulus categories

This study demonstrates that the human inferior temporal gyrus regions associated with mathematical cognition (ITG-math) are primarily engaged in processing numerosity across different visual stimulus categories rather than being selectively tuned to digits.

The Gist

The Brain's "Number Factory": A Simple Explanation

Imagine your brain is a bustling city. In the back part of this city, in a neighborhood called the Inferior Temporal Gyrus (ITG), there is a specific district that scientists have been arguing about for years.

Some researchers thought this district was a specialized "Digit Shop" that only cared about seeing numbers written as symbols (like 1, 2, 3). Others thought it was a "Math Factory" that cared about the concept of numbers, no matter what they looked like.

This new study acts like a detective, sending a team into that district to figure out what it actually does. Here is what they found, explained simply.

---

1. The Experiment: The "Alien vs. Astronaut" Game

To solve the mystery, the researchers set up a game for 17 volunteers. They showed them pictures on a screen and asked them to play two different games:

• The "Counting Game" (The Math Task): The volunteers had to press a button if they saw the same number of items twice in a row.
• The "Color Game" (The Control Task): The volunteers had to press a button if they saw the same color twice in a row.

The Twist: The items they were counting weren't just numbers. Sometimes they were Arabic Digits (like "5"). Other times, they were UFOs (aliens) arranged in groups of 1, 2, 3, 4, or 5.

Think of it like this:

• Scenario A: You see a "5" (digit) and then another "5". Do you press the button? (Counting Game)
• Scenario B: You see a group of 5 UFOs and then another group of 5 UFOs. Do you press the button? (Counting Game)
• Scenario C: You see a blue "5" and then a blue "5". Do you press the button? (Color Game)

2. The Big Discovery: It's About the Job, Not the Object

The researchers scanned the volunteers' brains while they played. Here is the "Aha!" moment:

• The "Digit Shop" Theory Failed: When they looked for brain cells that only liked seeing the symbol "5" and hated seeing 5 UFOs, they barely found any. It was like looking for a shop that only sells apples and finding out it sells oranges, bananas, and grapes too. Only about 35% of people had this specific "digit-only" zone.
• The "Math Factory" Theory Won: However, when they looked for brain cells that lit up during the Counting Game (regardless of whether it was digits or UFOs) compared to the Color Game, they found a massive, reliable zone in 92% of the people.

The Analogy:
Imagine a factory worker.

• If you ask, "Does this worker only like red bricks?" the answer is no. They work with blue bricks, green bricks, and alien-shaped bricks too.
• But if you ask, "Does this worker get excited when they are building a wall?" the answer is a resounding yes.
• The brain region (ITG-math) doesn't care if the "bricks" are numbers or aliens; it cares that the worker is doing math.

3. Left vs. Right: The Specialized Twins

The study also found that the two sides of this "Math Factory" work slightly differently, like a pair of twins with different personalities:

• The Left Side: This side is the Task Master. It screams, "We are doing math!" It doesn't really care what the objects look like. It's focused entirely on the job being done.
• The Right Side: This side is the Detail Oriented. It also knows we are doing math, but it also notices the details of the objects (like whether they are UFOs or digits). It's like a manager who knows the project is "Construction" but also keeps an eye on the specific materials being used.

4. Why This Matters

For a long time, scientists thought our brains had a special "number recognition" spot that only worked for the symbols we learn in school (1, 2, 3). This study suggests that's not quite right.

Instead, this part of the brain is a universal number processor. It's the part of your brain that helps you understand "how many" things there are, whether you are looking at:

• Digits on a clock.
• A flock of birds.
• A pile of coins.
• A group of aliens.

The Takeaway

The brain region in question isn't a "Digit Detector." It's a "Quantity Detector."

The researchers also created a new, efficient "map" (a localizer) that can find this specific brain zone in almost anyone in just 10 minutes. This is like giving future scientists a GPS to find the "Math Factory" quickly, so they can study how we learn math, why some people struggle with it, and how our brains handle numbers in the real world.

In short: Your brain doesn't just recognize the shape of a number; it recognizes the idea of a number, no matter what form it takes.

Technical Summary

1. Problem Statement

The functional role of the posterior inferior temporal gyrus (pITG) in mathematical cognition remains a subject of debate. Two primary hypotheses exist:

• Digit-Selectivity Hypothesis: Specific subregions (often termed "Visual Number Form Areas" or ITG-numbers) are selectively tuned to Arabic numerals (digits) regardless of the task.
• Stimulus-Independent Hypothesis: Regions in the pITG (termed ITG-math) are involved in mathematical processing (specifically numerosity) independent of the visual format or stimulus category (e.g., digits vs. non-symbolic arrays).

Previous studies have yielded conflicting results, with some identifying digit-selective regions and others finding only task-dependent activation. This study aims to disentangle the effects of visual stimulus features (digits vs. non-symbolic UFOs) and task demands (numerosity processing vs. color processing) to determine the true functional nature of pITG subregions.

2. Methodology

Participants:

• 17 healthy adults (11 females, mean age 27 ± 4.15 years) completed two separate fMRI sessions.
• Two participants were excluded after the first session; the final dataset included 17 participants.

Experimental Design (UFO-Paradigm):
The study employed a 2x2 factorial design manipulating Stimulus (Digits vs. UFOs) and Task (Numerosity vs. Color) orthogonally.

• Stimuli:
  • Digits: Arabic numerals (1–5) in three different fonts.
  • UFOs: Non-symbolic arrays of 1–5 alien figures in three different shapes.
  • Both categories were presented in five hues of blue.
• Tasks (1-back paradigm):
  • Numerosity Task: Participants pressed a button if two consecutive images had the same number of items (e.g., two 3s or two 3-UFOs).
  • Color Task (Control): Participants pressed a button if two consecutive images had the same color hue.
• Procedure: Two sessions were conducted. Each session consisted of four runs (~5 mins each), with 28 blocks per run. A cue indicated the task (alien hands for numerosity, palette for color).

Data Acquisition & Preprocessing:

• Scanner: Siemens 3T Prisma.
• Sequences: Gradient echo-planar imaging (EPI) for functional data (TR=1500ms, 2.5mm isotropic) and T1-weighted anatomical scans.
• Preprocessing: Motion correction, alignment to T1, no spatial smoothing. Analysis used the mrVista toolbox.
• Localizer: A separate color localizer (faces, houses, limbs, text, objects in color vs. grayscale) was used to define control regions in the ventrotemporal cortex.

Analysis Strategy:

• fROI Definition: Regions of Interest (ROIs) were defined in Session 1 and tested on independent data from Session 2 to avoid circularity.
  • ITG-math: Defined by contrasting Numerosity Task > Color Task.
  • ITG-numbers: Defined by contrasting Digits > UFOs.
• Statistical Tests: Repeated-measures ANOVAs (rmANOVAs) for univariate responses (Beta weights).
• Multivoxel Pattern Analysis (MVPA): Representation Similarity Analysis (RSA) and Winner-Takes-All (WTA) classifiers were used to decode task and stimulus information from distributed voxel patterns.
• Reliability: Spatial consistency across sessions was quantified using the Dice Coefficient (DC).

3. Key Contributions

1. Resolution of the "Digit vs. Math" Debate: The study provides strong evidence that pITG regions are primarily engaged in numerosity processing (task-dependent) rather than being strictly selective for digits (stimulus-dependent).
2. Development of a Robust Localizer: The authors propose and validate a specific experimental paradigm (UFO-paradigm) that reliably identifies "ITG-math" regions in ~92% of participants, addressing the inconsistency of previous localization methods.
3. Hemispheric Asymmetry in Encoding: The study reveals a functional dissociation between hemispheres:
   • Left ITG-math: Encodes task information (numerosity vs. color).
   • Right ITG-math: Encodes both task and specific stimulus features (digits vs. UFOs).
4. Spatial Consistency: Demonstrated that ITG-math locations are highly consistent across sessions within individuals, supporting the existence of a stable functional subregion.

4. Key Results

Behavioral Data:

• Participants performed with high accuracy (>90%) in all conditions.
• Reaction times (RTs) were faster for the Color task than the Numerosity task.
• An interaction was found: RTs were faster for Digits in the Numerosity task but faster for UFOs in the Color task.

fROI Identification & Activation:

• ITG-math (Task Selectivity): Contrasting Numerosity > Color successfully identified functional ROIs in 92% of participants (16/17 left, 15/17 right). These regions showed significantly higher activation during the numerosity task in independent data.
• ITG-numbers (Stimulus Selectivity): Contrasting Digits > UFOs failed to identify reliable regions, finding selectivity in only 35% of participants (7/17 left, 5/17 right).
• Stimulus Preference: Within the identified ITG-math regions, univariate responses were actually higher for UFOs than for Digits during the numerosity task, likely due to the higher cognitive load of processing non-symbolic numerosities.

Spatial Reliability:

• The Dice Coefficient (DC) for ITG-math overlap between Session 1 and Session 2 was significantly above chance in both hemispheres (Left: 0.17, Right: 0.31), confirming stable localization.

Multivariate Analysis (MVPA/RSA):

• Left Hemisphere: Strong correlation with the Task Model (r=0.87), but no correlation with the Stimulus Model.
• Right Hemisphere: Significant correlation with both the Stimulus Model (r=0.65) and the Combined Task/Stimulus Model (r=0.66), but not the Task Model alone. This suggests the right hemisphere retains more visual feature information alongside task processing.

Whole-Brain Analysis:

• The pITG activation for numerosity was part of a broader network including the bilateral Intraparietal Sulcus (IPS), left Supramarginal Gyrus (SMG), and left Precentral Sulcus (PCS).

5. Significance

• Theoretical Impact: The findings challenge the notion of a dedicated "number form area" strictly for digits. Instead, they support a model where the pITG contains a domain-specific region for numerosity processing that operates across visual formats (symbolic and non-symbolic).
• Methodological Utility: The study provides a validated, efficient (10-minute data sufficient) localizer paradigm for future research. This allows researchers to reliably target ITG-math for studies on mathematical cognition, dyscalculia, or neural plasticity.
• Neural Mechanism: The discovery of hemispheric asymmetry (Left: Task; Right: Task + Stimulus) offers new insights into how numerical information is processed and integrated across the brain, suggesting a potential division of labor where the right hemisphere may initially process non-symbolic magnitude before integration in the left.
• Future Directions: The authors note that while ITG-math is robustly identified, the elusive "ITG-numbers" (digit-selective) may require different stimuli, passive viewing tasks, or specialized imaging parameters to detect, suggesting the field must refine its search criteria for digit-specific regions.