Brains that don’t see in greyscale first over-rely on colours: Project Prakash study
Premium
Project Prakash treats and rehabilitates blind children in India. These children helped researchers shed light on how the brain learns to see
The world has many colours and being able to see them is a great source of joy. But a newborn baby mostly sees the world in black and white. The photosensitive cone cells in the child’s eyes don’t mature until they’re around four months old. In this time, the brain uses other visual cues to make sense of the world.
In May, a team of Indian and U.S. researchers reported in the journal Science that this delay in developing colour vision is actually important for overall vision development.
“We are able to explain why normal visual development happens the way it does,” study co-lead Priti Gupta, a cognitive neuroscientist, told The Hindu.
What are colours and how do people understand them? | Explained
Dr. Gupta leads the research team in India for ‘Project Prakash’, a U.S.-based initiative of IIT Delhi, Dr Shroff’s Charity Eye Hospital in New Delhi, and the Massachusetts Institute of Technology.
Project Prakash treats and rehabilitates blind children in India. These children helped the researchers shed light on how the brain learns to see.
Importance of colour vision
Humans don’t need colour vision to recognise objects but colours can provide adaptation and survival advantages, Dr. Gupta said.
“If you look at food, you can recognise it without colour. But if something has gone bad, then you need to see the colour of that thing to understand.”
Yet there is more to colour vision than meets the eye.
“They’re very mesmerised by colours,” Dr. Gupta said of the children at Project Prakash. They often described objects around them with their colour. “Their reliance on colours is a little more than what normal children have.” This observation gave the researchers an idea: “to show them some things without colour”.
They brought together a group of visually impaired children and young adults between 8 and 26 years of age and asked them to identify objects in images — a tree, a bus, chicken, a stack of books, etc. — first in grayscale and then in colour. In another test, they asked the group to determine which of two discs presented to them had a lighter hue while the researchers adjusted the colour.
The children could recognise colour images and discs quite well — even those who were barely two days out of eye surgery. But they had a tough time recognising black and white images.
Children without any visual impairment had trouble neither with colour nor grayscale images, on the other hand.
The researchers realised the issue wasn’t the colour vision of the visually impaired children. Instead, it was the abnormal way in which their vision had developed: to rely too much on colour.
Members of the Project Prakash team in India (from left): Abhishek Kumar, Rakesh Kumar, Ajay Chawariya, Priti Gupta, Shakila Bi, Suma Ganesh, Ranupriya, Naviya Lall, and Dhun Verma.
| Photo Credit:
Himanshu Kumar
Mimicking visual development
Normally, a child first understands the world in grayscale. But the first time the children at Project Prakash experienced normal vision, their eyes had developed enough to see colours as well, so they skipped the grayscale phase. Their brain processed black and white images differently as a result.
To understand the effects of this issue, the researchers needed a proxy to the brain that they could tweak to learn in response to different visual stimuli. They set up a deep convolutional neural network (CNN) — a computer program that processes information the way neurons in the brain’s visual cortex do. Engineers have previously used deep CNNs in image recognition software.
“They are not perfect models. But they are the best models people have at this point of time,” Dr. Gupta said.
They trained four CNNs, one each on colour and grayscale images in a particular order: grey-grey, colour-colour, colour-grey, grey-colour.
They found the grey-CNN recognised both greyscale and colour images better than any of the other models.
The colour-colour model — which most mimicked visual development among Project Prakash’s children — fared worse at identifying greyscale images.
The researchers attributed this to the colour-colour model’s overreliance on colour cues when examining images because its training data was composed solely of colour images. The grey-colour model had learnt enough cues from the greyscale images and was thus better able to recognise colour images.
These findings suggest why the children too performed poorly in colour tests. “They already start out with very clear vision as compared to a normal, developing baby,” Dr. Gupta said.
How the colour blind see coloured medication
Optimising visual development
It’s fascinating that the brain develops object recognition and colour perception at different times, study co-author Dr Suma Ganesh and Project Prakash’s clinical lead said.
Dr. Ganesh is an eye surgeon, pediatric ophthalmologist, and medical director at Dr. Shroff’s Charity Eye Hospital in New Delhi.
“We never thought that if one [perception] is d, it has a lot of impact on the other.” she said.
It’s unclear whether this particular study would bring about any rehabilitation benefits though.
“At the moment, I wouldn’t say we have done anything much on the rehabilitation side,” Dr. Gupta said. “But all of this knowledge that’s coming to us will eventually be used for better treatments.”
But she added that there are ideas that can help. For example, children could also be made to experience a room deprived of colour, simulating a black and white or a greyscale environment, for a few hours at a time. “Maybe that helps to fine tune their cells to actually generalise better,” she said.
However, she admitted such ‘tests’ could also have ethical concerns and that their efficacy would have to be checked with animal models first.
Karthik Vinod is interning with The Hindu.
- Telegram
SEE ALL
PRINT
research
/
opthalmology
/
Artificial Intelligence