Behavioural stochastic resonance across the lifespan

Abstract

Stochastic resonance (SR) is the phenomenon wherein the introduction of a suitable level of noise enhances the detection of subthreshold signals in non linear systems. It manifests across various physical and biological systems, including the human brain. Psychophysical experiments have confirmed the behavioural impact of stochastic resonance on auditory, somatic, and visual perception. Aging renders the brain more susceptible to noise, possibly causing differences in the SR phenomenon between young and elderly individuals. This study investigates the impact of noise on motion detection accuracy throughout the lifespan, with 214 participants ranging in age from 18 to 82. Our objective was to determine the optimal noise level to induce an SR-like response in both young and old populations. Consistent with existing literature, our findings reveal a diminishing advantage with age, indicating that the efficacy of noise addition progressively diminishes. Additionally, as individuals age, peak performance is achieved with lower levels of noise. This study provides the first insight into how SR changes across the lifespan of healthy adults and establishes a foundation for understanding the pathological alterations in perceptual processes associated with aging.

Keywords: Ageing; Neural noise; Psychophysics; Stochastic resonance; Visual perception.

Fig. 1

Frequency distribution of the age of the participants. Histogram represents the frequency distribution of age in the dataset, with a density curve (grey line) overlaid. Green line indicates the mean age, while the red line represents the median age. Bin width for the histogram is set to 4 units. The plot provides insights into the central tendency and distribution of age in the examined population

Fig. 2

Statically, two examples of the configuration of moving dots presented throughout the experiment. A configuration of moving dots with a percentage of coherently moving dots (left). All the dots move randomly (right)

Fig. 3

Trial sequence consisted of a 133-ms initial interval featuring coherent (or random) motion, followed by a 1000-ms poststimulus gap. Subsequently, there was a second 133-ms interval with random (or coherent) motion, another 1000-ms poststimulus gap and, finally, the participant's response. During the practice block, the response was followed by visual feedback. Trial-by-trial instructions were exclusively presented during practice trials. In the experimental trials, a red circle replaced the instructions, signalling participants when to provide their response

Fig. 4

Age-related changes in threshold for coherent RDK. The plot shows how the ln(threshold) for detecting coherent motion, expressed as the natural logarithm of the percentage of coherent dots out of the total dots, varies with age for staircases with 100 and 400 dots. Linear regression lines, shaded 95% confidence bands, and individual data points are included. Red line corresponds to the staircase with 100 dots. Green line represents the staircase with 400 dots

Fig. 5

The relationship between accuracy, dot numerosity, and age. Lines represent predictions of the fourth-degree model. Although the age variable in the model is inherently continuous, for visualization, we have represented it through four age levels. The x-axis represents dot numerosity on a logarithmic scale. The coloured bands represent the 95% confidence intervals

Fig. 6

The relationship between accuracy, dot numerosity, and age group as a categorical two-level factor. Dots represent average accuracy for each group. Lines depict the predictions of the fourth-degree model. X-axis denotes dot numerosity on a logarithmic scale. Coloured bands represent the 95% confidence intervals