Revealing Different Cognitive Mechanisms of the Ebbinghaus Illusion Through Neural Oscillations

Academic Background

Human perception of size in vision is not entirely faithful to the physical world and is highly dependent on context. For example, when an object is surrounded by several smaller objects, it appears larger than when surrounded by larger objects. This phenomenon is known as the Ebbinghaus illusion. Visual illusions provide a unique perspective for understanding the mechanisms of conscious experience in the visual world. The Ebbinghaus illusion can be explained by two cognitive mechanisms: low-level contour interaction and high-level size contrast. Low-level contour interaction is a sensory interaction that occurs at the level of contours or features, leading to perceptual distortion when a figure is surrounded by other figures. This mechanism may be related to local circuits within the primary visual cortex (V1). High-level size contrast, on the other hand, relies on cognitive size comparison between the central target and surrounding inducers, resulting in perceptual enhancement of their size difference. The high-level size contrast mechanism of the Ebbinghaus illusion is thought to require feedback connections from higher visual areas, particularly feedback projections from the right parietal cortex to the occipital region. However, the neural mechanisms of these two theories remain largely unexplored.

Literature Source

This study was conducted by Lihong Chen, Yi Jiang, and other authors from the State Key Laboratory of Brain and Cognitive Science, Institute of Psychology, Chinese Academy of Sciences, and several other research institutions. The paper was submitted on November 2, 2024, and published in Neurosci. Bull. (March 15, 2024).

Research Process

This study used stereoscopic illusion presentation and electroencephalography (EEG) techniques to investigate changes in stimulus range and potential neural mechanisms in the Ebbinghaus illusion by manipulating the spatial relationship between the central target and surrounding inducers. The study included multiple experimental steps:

Experimental Procedure

• Subjects: 30 right-handed healthy volunteers (average age 23.3 years).
• Experimental Design and Stimuli: Experiment stimuli were configured using Matlab software package and displayed on a 20-inch cathode-ray tube monitor. The relative depth position of the central target was manipulated by adjusting the horizontal disparity of the surrounding circles.

Experiment 1

Task 1: Observers needed to judge the position of the target relative to the inducers.

Task 2: Adjust the size of a comparison circle to match the target.

Results: When the target was presented on a different depth plane from the inducers, the illusion effect was significantly reduced.

Experiment 2

Task: Observers viewed the stimuli directly with both eyes without using a stereoscope and only needed to perform the size matching task.

Results: The illusion effect was significantly reduced in high and low disparity conditions under different disparity conditions.

Experiment 3

Task: Present four inducers on different depth planes, and observers needed to judge depth and match size.

Results: The illusion effect disappeared on visually different depth planes.

Experiment 4

Task: Observers viewed stimuli through a stereoscope, judged the position of the target relative to the inducers, while ERP data was recorded.

Results: EEG data showed that when there were large inducers around the target, the target evoked larger N1 and P2 amplitudes. Under depth conditions, larger perceived size of the target was associated with lower early α-wave power, while in the zero disparity condition, larger perceived size was associated with higher late β-wave power.

EEG Data Recording and Analysis

Recording: Continuous EEG data was recorded using a 64-electrode cap.

Analysis: EEG data was analyzed using EEGLAB, filtering out activity below 1Hz and above 40Hz. Event-related spectral perturbation (ERSP) was baseline-corrected after the appearance of illusory figures. EEG data was transformed into the time-frequency domain using the Fast Fourier Transform (FFT) method, extracting α-wave (8–13 Hz) and β-wave (14–25 Hz) amplitudes separately.

Research Results

Depth Cues Reduce Size Illusion Effect

Under depth cue conditions (Experiment 1), the illusion effect was significantly reduced. In high and low disparity conditions, the illusion effect was significantly lowered. Experiment 3 showed that when four inducers were displayed on different depth planes, the illusion effect disappeared.

Neural Oscillations Associated with the Ebbinghaus Illusion

Early α-wave power decrease in the centro-parietal region was negatively correlated with the illusion effect. In the zero disparity condition, larger perceived size produced higher β-wave power in the parieto-occipital region in a later time window (200–300 ms), and β-wave power was positively correlated with behavioral illusion effects. The results suggest that induced α-wave and β-wave oscillations are associated with low-level contour interaction and high-level size contrast, respectively.

Conclusion

This study provides neurophysiological evidence supporting two cognitive mechanisms of the Ebbinghaus illusion by revealing the dynamic support of neural oscillations in different frequency bands for various aspects of visual processing. Early α-wave power is associated with low-level contour interaction, while relatively later β-wave power is associated with high-level size contrast. This finding suggests that local circuits in the primary visual cortex and neural oscillations in different frequency bands in the parieto-occipital region dynamically support each other in various aspects of visual perception.

Research Significance

This study not only reveals the neural basis of low-level and high-level cognitive mechanisms in visual perception but also provides potential applications for future visual enhancement and perceptual compensation through neural oscillation modulation. Additionally, the experimental methods and data analysis techniques used in the study provide new research paradigms for the field of visual neuroscience.