Improving Vision Using Noise

Overview

• Stochastic Resonance (SR) is a phenomenon where the detection of a weak stimulus is enhanced with the help of optimal levels of noise.
• Itzcovich et al. (2017) explored the effect of SR in individuals with various visual impairments.
• They found that visual detection was improved at optimal amounts of noise, and worsened when there was too much noise present!
• The results indicate that noise could be used in visual prosthetics for various impairments, and the study contributes greatly to our current understanding of how something as annoying as noise can be used as a benefit in a clinical setting!

In my previous post, I discussed how noise can enhance the performance of a system via Stochastic Resonance (SR). If you missed it, I would recommend you to click the link here so you can catch up on the basics. But just as a quick refresh, SR is a phenomenon in which the detection of a weak stimulus is enhanced with the help of optimal levels of noise. In simple terms, a stimulus that was previously difficult to detect for an individual, is now easy to detect with the help of optimal levels of noise. The noise used is generally a smoothly presented background visual noise (noise without the auditory component). An example of visual noise is provided on the right.

In this post, I will cover a research paper that was published recently in which, the SR effect was explored in a more practical experiment. This post will also serve as evidence which illustrates the efficacy of SR.

Again, I would strongly recommend you read the post Noise can be good for you! to cover the basics.

Improving vision in the visually impaired

Itzcovich et al. (2017) explored the effect of SR in individuals with various visual impairments. Whilst there are many papers that show the SR effect, this paper was one of the first to look at the SR effect in the visually impaired. There were 14 participants in this study, with demographics summarised in the table below, along with the associated visual impairment.

Age	Sex	Visual disorder
50	M	Optic atrophy
28	F	Optic atrophy
49	F	Degenerative myopia
53	M	Retinitis Pigmentosa
49	M	Retinitis Pigmentosa
33	F	Retinitis Pigmentosa
45	M	Retinitis Pigmentosa
34	F	Optic atrophy
25	F	Retinitis Pigmentosa
51	F	Macular degeneration
44	F	Retinitis Pigmentosa
27	F	Retinitis Pigmentosa
22	M	Degenerative myopia
30	F	Degenerative myopia

For the visual detection task, participants had to detect the letters C, D, H, K, N, O, R, S, V, and Z when they were displayed on a computer monitor. The experimenters ensured that pairs of visually similar letters such as O-C, H-N, or R-K, were never presented in close sequence to avoid uncontrolled ambiguity or guessing. Gaussian white noise was used as added background noise in this experiment. For each of the ten letters, six levels of noise were presented. The noise levels were σ = 6, σ = 12, σ = 18, σ = 30, σ = 60, and σ = 90. Note: ‘σ ‘ symbol denotes sigma.

The image below provides an example for the no noise condition (original condition) and noise condition (σ = 6 to σ = 90).

In their experiment, they found that at baseline (zero noise), the fraction of recognised letters among participants ranged between 0 and 0.3. However, when noise was added, this range was increased. The maximum recognition rate ranged between 0.2 and 0.8 at noise levels between σ = 6 and σ = 30. The performance degraded, or the letter recognition rates decreased in all participants, at noise levels above σ = 30. The nature of the results was plotted graphically by the researchers and is illustrated below. This inverted-U like graph depicts the signature function of Stochastic Resonance (SR).

These results show that it is possible to develop a prosthetic device, such as glasses that can implement noise to aid people with certain visual impairments. Obviously, more research is needed to replicate these findings and potentially test this SR effect in a more practical setting than just a computer based visual detection task. It is also important to note that there are many factors that influence SR, such as the noise levels within the brain itself and therefore, developing such noise-based visual prosthetics may bring certain challenges. Hence, future research needs to look at a more practical approach which takes into account individual differences. Additionally, future research can also look at how participants perform when other stimuli such as faces, objects, and images are implemented. This can also provide insight on the practicality of the SR phenomenon.

Nevertheless, this experiment provided potentially the first clinical evidence of SR, in which the detection of letters, which the participants struggled to detect at baseline condition, was enhanced by the addition of optimal levels of visual noise. This research contributes greatly to our current understanding of how something as annoying as noise can be used as a benefit in a clinical setting. Kudos to Itzcovich and colleagues!

I recommend you to check out the full research paper Stochastic resonance improves vision in the severely impaired.

Featured Image Credit sdecoret / Shutterstock

More From The NeuroBlog

• How can Neurofilament light concentrations help in Diagnosis?
• Execessive Neural Noise in Developmental Dyslexia?
• Debunking Popular Neuromyths: Do You Use Your Entire Brain?
• Neurology and Mental Health: The Impact of COVID-19
• Neuromyths: What are they and what can we do about them?

• Instagram
• Facebook
• LinkedIn

Latest from Instagram