Cognition does not automatically influence perception: Evidence from neural encoding of colours belonging to different categories

This study challenges the view that cognition automatically influences early perception by demonstrating that neural responses to color changes, previously attributed to categorical processing in speakers with multiple blue terms, are actually driven by low-level contrast adaptation rather than linguistic categories.

The Gist

The Big Question: Does Language Rewire Your Brain?

Imagine your brain is a high-tech security camera system. For decades, scientists have debated a fascinating question: Does the language you speak change how that camera sees the world?

This idea is called the Whorfian hypothesis. It suggests that if your language has two different words for "light blue" and "dark blue" (like Russian does), your brain might automatically treat those two colors as totally different categories, even before you consciously think about them. It's like having a security filter that instantly flags "light blue" as a "different intruder" than "dark blue," whereas someone speaking English (who just calls them both "blue") might not notice the difference as quickly.

The Previous "Proof" (The Ghost in the Machine)

A few years ago, a famous study claimed to find the "smoking gun" for this theory. They used EEG (a cap with electrodes that reads brain waves) to watch people's brains react to colors. They found that Greek speakers (who also have two words for blue) showed a specific, quick brain spike when switching between light and dark blue. This spike happened so fast (in a fraction of a second) that it seemed to prove language was hacking the brain's early visual system.

It was like seeing the security camera flash a red warning light before the guard even realized a person had walked in.

The New Study: "Wait, Let's Check the Wiring"

The authors of this new paper decided to play detective. They thought, "Maybe that red warning light wasn't about language at all. Maybe it was just about the brightness or contrast of the colors."

They ran three new experiments to test this. Think of it as a series of stress tests for the brain's security system.

Experiment 1: The Russian Replication

They took the exact same test but used Russian speakers (who also have two words for blue: goluboj and sinij).

• The Setup: They flashed light blue and dark blue, then light green and dark green.
• The Result: The "language effect" vanished. The brain didn't show that special spike for blue just because they had two words for it. The brain waves for blue and green looked almost identical.
• The Analogy: It's like checking a smoke detector. The old study said, "It beeped because it smelled smoke!" The new study says, "Actually, we just opened the window, and the wind blew the dust sensor. It wasn't smoke; it was just a draft."

Experiment 2: The "Warm vs. Cool" Test

They tested English speakers. In English, the "cool" colors (blue/green) are just one category, but the "warm" colors (red/pink, yellow/brown) have many distinct names.

• The Prediction: If language drives the brain, English speakers should have a huge brain spike for the "warm" colors because they have more words for them.
• The Result: Nope. The brain didn't care about the words. It only cared about contrast. When the colors were bright and high-contrast, the brain reacted. When they were dull or low-contrast, the brain stayed quiet.
• The Analogy: Imagine a bouncer at a club. The old theory said the bouncer checks your ID (your language) to let you in. The new study shows the bouncer is actually just checking if you are wearing a bright neon jacket (high contrast). If you are wearing neon, you get in. If you are wearing a dull grey suit, you don't, regardless of what your ID says.

Experiment 3: The "Control Freak" Test

Finally, they isolated the variables. They changed the colors, the brightness, and the saturation (how "rich" the color is) one by one.

• The Finding: The brain's "mismatch" signal (the vMMN) only fired when there was a difference in contrast or brightness. It didn't fire just because the color changed from "Red" to "Green" if the brightness stayed the same.
• The Conclusion: The brain isn't reacting to the name of the color; it's reacting to the physical energy of the light hitting the eye.

The Real Culprit: "Visual Adaptation"

So, what was actually happening in the brain? The authors suggest it's a phenomenon called Contrast Adaptation.

The Analogy of the Noisy Room:
Imagine you are sitting in a room with a fan humming loudly (the "standard" stimulus). Your brain gets used to the hum and ignores it. Then, the fan suddenly stops or changes pitch (the "deviant"). Your brain jumps: "Wait, something changed!"

The old study thought the brain jumped because the color changed.
The new study shows the brain jumped because the loudness (contrast) of the fan changed. If the fan was always loud, and then got quiet, the brain reacted. If the fan was always quiet, and then got loud, the brain reacted. But if the fan stayed the same volume and just changed color, the brain barely noticed.

Why This Matters

This paper is a major "plot twist" for cognitive science.

1. Perception is Bottom-Up: It suggests that our early vision is driven by physics (light, contrast, brightness) rather than by our thoughts or language. The brain is a machine that processes light first, and language second.
2. The "Ghost" was a Glitch: The famous evidence that language shapes perception might have been a statistical fluke or a misunderstanding of how contrast works.
3. Predictive Coding: It challenges how we think the brain predicts the future. It suggests the brain's "prediction error" signals are actually just the brain getting used to the background noise (adaptation), not necessarily a complex linguistic calculation.

The Bottom Line

The paper concludes that cognition (thinking/language) does not automatically penetrate perception (seeing).

Your brain doesn't see the world through the lens of your vocabulary. It sees the world through the lens of light and contrast. If you speak Russian and see a light blue sky, your brain doesn't instantly scream "Different Category!" just because you have a word for it. It just sees a patch of light. The "magic" of language happens later, when you consciously think about what you are seeing, not in the first split-second of raw perception.

Technical Summary

1. Problem Statement and Background

The study addresses a central debate in cognitive science: the Whorfian hypothesis (linguistic relativity), which posits that language categories can directly shape early, pre-attentive perceptual processes. Specifically, the authors investigate the claim that linguistic color categories alter the neural encoding of color before conscious attention is engaged.

• The Preceding Evidence: The primary evidence for this claim comes from a seminal 2009 study by Thierry et al., which used Electroencephalography (EEG) to show that Greek speakers (who have two basic terms for blue: ghalazio and ble) exhibited an enhanced Visual Mismatch Negativity (vMMN) when distinguishing between light and dark blue compared to light and dark green. The vMMN is an event-related potential (ERP) component peaking at 150–240 ms, traditionally interpreted as a marker of pre-attentive predictive coding and feature-change detection.
• The Critique: Recent literature suggests that the vMMN is highly susceptible to contrast adaptation (repetition suppression) rather than purely cognitive or categorical processing. The authors hypothesize that the original findings by Thierry et al. may have been confounded by low-level physical properties (luminance and chromatic contrast) rather than linguistic categorization.

2. Methodology

The authors conducted a three-experiment study to systematically re-evaluate the evidence for early categorical effects on color perception.

• Participants:
  • Experiment 1: 27 native Russian speakers (who, like Greeks, have two basic terms for blue: goluboj for light blue and sinij for dark blue).
  • Experiment 2: 36 native English speakers (who have distinct basic terms for "warm" colors like red/pink and yellow/brown, but fewer distinct basic terms for "cool" colors like blue/green).
  • Experiment 3: 25 participants (mixed language background) in a preregistered follow-up.
• Paradigm: All experiments utilized a classical oddball paradigm. Participants viewed streams of circular stimuli where a "standard" color was repeated 3–5 times, followed by a "deviant" color. Participants were instructed to ignore color changes and only respond to shape changes (squares vs. circles), ensuring the color processing was task-irrelevant and pre-attentive.
• Stimuli and Manipulations:
  • Exp 1: Close replication of Thierry et al. using Russian speakers with light/dark blue and green stimuli.
  • Exp 2: Conceptual replication with English speakers, manipulating "warm" (red/pink, yellow/brown) vs. "cool" (blue/green) regions. Crucially, this experiment introduced a small, controlled difference in luminance contrast between light and dark shades to test adaptation effects.
  • Exp 3: Systematic manipulation of three perceptual dimensions: Hue (Red vs. Green), Saturation (Chromatic contrast), and Lightness (Luminance contrast). This allowed the authors to isolate whether vMMN is driven by categorical boundaries or physical contrast.
• EEG Analysis:
  • Data were recorded using 64-channel EEG systems.
  • Preprocessing included artifact rejection (FASTER/ADJUST toolboxes), filtering (0.1–40 Hz), and baseline correction.
  • vMMN Calculation: Derived by subtracting the ERP response to standard stimuli from the response to deviant stimuli.
  • Statistical Approach: Linear Mixed Effects Models (LMEMs) and Bayesian t-tests (calculating Bayes Factors, BF) to determine the reliability of effects against zero and the influence of specific predictors (hue, lightness, saturation, deviancy).

3. Key Results

Experiment 1 (Russian Speakers):

• Failure to Replicate: The study failed to replicate the enhanced vMMN for blue categories found in the original Greek study.
• Lightness Dominance: The vMMN was not driven by the categorical distinction between sinij and goluboj. Instead, the effect was driven entirely by lightness.
  • Robust vMMN negativity was observed only for light deviants (following dark standards).
  • No reliable vMMN was found for dark deviants.
  • Statistical models showed a significant interaction between deviancy and lightness, but no effect of color category (blue vs. green).

Experiment 2 (English Speakers):

• No Categorical Effect: English speakers did not show enhanced vMMN for "warm" colors (which have more basic terms) compared to "cool" colors.
• Contrast Adaptation: The results mirrored Experiment 1. Significant vMMN effects were observed only when there was a difference in luminance contrast between the standard and deviant.
  • Lighter colors elicited a robust vMMN; darker colors did not.
  • An early positive deflection (~110 ms, P1) was observed for darker, higher-contrast stimuli, consistent with OFF-pathway luminance processing rather than categorical processing.

Experiment 3 (Systematic Feature Manipulation):

• Hue: No reliable effect of hue (Red vs. Green) on vMMN was found (BF = 0.197), even though these are distinct basic categories.
• Saturation: Effects were asymmetrical; vMMN was only elicited when the deviant was more saturated than the standard. Desaturated changes did not elicit a robust vMMN.
• Luminance Contrast: Strong evidence (BF > 10) that vMMN amplitude is modulated by luminance contrast.
• Conclusion: The vMMN is driven by contrast adaptation mechanisms (specifically to luminance and chromatic contrast) rather than the categorical distinctness of the colors.

4. Key Contributions

1. Refutation of Early Categorical Penetrability: The study provides robust, high-powered evidence against models proposing that linguistic color categories automatically penetrate early, pre-attentive visual processing (the "top-down" influence).
2. Reinterpretation of vMMN: The authors demonstrate that the vMMN in color oddball tasks is largely a signature of contrast adaptation and repetition suppression, not necessarily a marker of predictive coding based on abstract categories.
3. Methodological Correction: The paper highlights that previous studies (including Thierry et al.) likely confounded categorical effects with physical properties (lightness/contrast). By controlling for these low-level variables, the "categorical" effect disappears.
4. Asymmetry in Adaptation: The study identifies a novel asymmetry in how the brain adapts to chromatic contrast (vMMN only for higher-contrast deviants) versus luminance contrast.

5. Significance

• Cognitive Science: The findings strongly support the view that perception is largely encapsulated from cognition at early stages. Language categories do not automatically alter the neural encoding of color features before attention is engaged.
• Neuroscience & Predictive Coding: The results challenge the interpretation of vMMN as a pure index of predictive coding violations based on high-level categories. Instead, they suggest that early prediction errors in the visual domain may be rooted in sensory adaptation to physical stimulus properties (contrast).
• Clinical Implications: Since vMMN is used as a biomarker for predictive coding deficits in conditions like psychosis, this study suggests that researchers must rigorously control for low-level contrast confounds to ensure that observed deficits are truly cognitive/predictive rather than sensory adaptation artifacts.
• Future Directions: The authors propose that categorical effects on color perception likely require task-dependent attention or working memory engagement (e.g., comparing a color to a linguistic standard) rather than occurring automatically in passive viewing.