Simulations Show Musk’s Vision System Performs Poorly

Elon Musk’s ambitious eye implants, which aim to surpass normal human vision, are unlikely to achieve this goal. Scientists using ‘virtual patients’ have highlighted the limitations of this technology, showing that even the best engineering cannot fully replicate human neurophysiology in restoring sight.

Elon Musk’s ambitious eye implants, which aim to surpass normal human vision, are unlikely to achieve this goal. Scientists using ‘virtual patients‘ have highlighted the limitations of this technology, showing that even the best engineering cannot fully replicate human neurophysiology in restoring sight.

“Blindsight is the next product following Telepathy,” he posted. “I should note that the Blindsight implant is already functioning in monkeys. Initially, the resolution will be low, similar to early Nintendo graphics, but it may eventually surpass normal human vision. (Additionally, no monkey has died or been seriously harmed by a Neuralink device!)“

Musk has claimed that Blindsight will allow people without eyesight or who have lost their eyes to ‘see‘ by targeting the brain’s processing of optical information. This would involve implanting millions of tiny electrodes in the visual cortex, the area at the back of the brain responsible for processing and interpreting visual information from the eyes.

Researchers Highlight Fundamental Flaws in Cortical Implants Like Blindsight

However, researchers at the University of Washington (UW) argue that there are fundamental flaws in the design of cortical implants like Blindsight. They believe these implants underestimate the complexities of human eye-brain communication. Using detailed computational modeling simulations, referred to as virtual patients, UW researchers have demonstrated that these implants are unlikely to ever “exceed” normal vision.

The core issue lies in the limitations of these electrodes and their ability to stimulate the neurons required to artificially recreate vision. This complex process depends on generating a multitude of intricate neural codes needed for the brain to properly process visual information.

Achieving Human Vision Requires Complex Mapping of Visual Cortex Cells

“To achieve typical human vision, each cell in the visual cortex would need an electrode with the correct code,” said Ione Fine, a psychology professor at UW. “This is highly complex since each cell has a unique code. Mapping out every single cell would take years.”

Although 45,000 electrodes might be compared to 45,000 pixels on a screen, neurons relay information about overlapping “receptive fields.” A single spot of light stimulates many interconnected neurons, creating a significant challenge in replicating this with electrodes.

Simulations Reveal Limits of Treating Electrodes as Pixels

“Engineers often consider electrodes to be like pixels,” Fine stated, “but that’s not how biology functions. Our simulations, based on a basic model of the visual system, aim to provide insight into how these implants will perform. These simulations differ significantly from the intuition an engineer might have when thinking in terms of pixels on a computer screen.”

To illustrate this, the researchers created various simulations, including a movie of a cat shown at 45,000 pixels compared to how it would appear to a patient with 4,500 electrodes in their visual cortex. The implant models were based on data from existing studies on cortical implants similar to Blindsight’s concept. Although the electrodes could interpret a visual image, the cat appeared extremely blurry and hard to recognize beyond its basic shape.

In these videos, the researchers simulated two different electrode array configurations to show how the cat video would be perceived by someone with cortical implants.

Researchers Warn That Technology May Fall Short of Musk’s Vision

The researchers observed that while this approach would represent an improvement for someone completely blind, they warn it may never achieve the level of vision Musk envisions.

“Someone might eventually make a breakthrough that offers a ‘Rosetta Stone’ for this problem,” Fine said. “It’s also possible that individuals could develop some plasticity to better use an incorrect code. However, my research and that of others indicate there is no current evidence that people can significantly adapt to an incorrect code.”

Without replicating the necessary neural codes, no amount of engineering can make this technology approach adequate human vision. The researchers emphasize that this is a crucial consideration when evaluating the feasibility of biotech like Blindsight.

“Many people lose their sight later in life,” Fine noted. “At 70 years old, adapting to life as a blind individual is very challenging, and there’s often a high rate of depression and a strong desire to regain sight. While blindness itself doesn’t make people vulnerable, losing sight later in life can create vulnerabilities. Thus, when Elon Musk claims that this will surpass human vision, it’s a potentially dangerous statement.”

Read the original article on: New Atlas

Read more: