top of page
Medical Technology

Can a human brain be linked to a computer?

Scientists in the US have succeeded in mapping the three-dimensional structure of the network of neurons in one cubic millimetre of mouse brain- a feat that would require two petabytes of storage. The human brain contains approximately 100 billion neurons, which is one million times the number of neurons found in a cubic millimetre of mouse brain.

 

The researchers counted over 100,000 neurons and over a billion connections between them within this small cube of brain tissue. To find all the neurons and reconstruct the neural network, researchers had to slice the mouse brain 25,000 times. The issue is that the amount of data to store would kill any single computer. Memory and experiences that would have defined people later would be lost if they tried to store their minds too early. Using a computer too late may result in the accumulation of a mind with dementia, which would not work so well.

 

Human tissue would have to be cut into zillions of thin slices using techniques compatible with dying and cutting. Local electrical changes that travel down dendrites and axons allow neurons to communicate with one another. However, when reconstructing the 3D structure, this may not be possible. After we die, our brains undergo significant chemical and anatomical changes. At the age of 20, they begin to lose 85,000 neurons per day due to apoptosis, or programmed cell death. Many memories that would have shaped a person later would be lost if he or she tried to store their mind too early. There are numerous steps involved in developing a computer capable of storing and processing human-level intelligence.

 

It may be impossible for an artificial intelligence to produce sensations and actions identical to those provided and produced by your biological body. Bots are susceptible to hacking and hardware failure. Connecting sensors to the AI's digital mind would also be difficult.

By Jeevana Thavarajah 

Related article: The Evolution of AI

bottom of page