Utilizing artificial intelligence (AI) in combination with mind-reading technology is essential for recreating the sensations we experience in our dreams.
In a well-known Japanese study, the initial steps of this method were showcased in 2023. Researchers employed a functional magnetic resonance imaging (fMRI) scanner to monitor the brain activity of participants during sleep and utilized machine learning to identify recognized objects like keys, individuals, and chairs from that activity.
Nevertheless, this study concentrated on sleep onset, the first two stages of sleep where visual imagery (hallucinations) occur, and did not explore dreams at all.
They adopted this approach so participants could articulate what they observed upon waking.
To replicate dreams, we need an extensive database of detailed fMRI information from dreaming volunteers to educate a large-scale AI. Participants should possess exceptional recall abilities to describe their dreams vividly, which will help determine the accuracy of the predictions.
Recording dreams in this level of detail poses a significant challenge, and establishing a reliable method to generate such data remains uncertain.
Nonetheless, progress has already been made in related areas, with research studies producing vast datasets of fMRI brain activity from conscious participants watching videos, listening to spoken language, and reading text.
By employing AI trained on these datasets, we can already predict what people are viewing or reading while awake.
Assuming there is enough data to develop such an AI in a few years, as well as portable fMRI machines that allow for dreaming individuals to be monitored while asleep, the required methodologies to exhibit results will already be in place.
Generative AI such as OpenAI’s Sora and Google DeepMind’s Lumiere can already generate captivating video sequences. Utilizing dream analysis AI, when you provide a textual depiction to the generative AI, you receive a video illustrating the dream sequence.
However, it’s important to note that these AIs are not actually reading minds, but rather matching brain activity patterns with images that may have been previously seen. The generative AI cannot validate if the video accurately represents the dream—it simply pieces together images and possibly adds a rudimentary narrative.
Though the end product may strikingly resemble a dream with many familiar elements, it does not provide an exact replica, similar to how the movie Cast Away featuring Tom Hanks only loosely mirrors the true story of Jose Salvador Alvarenga, a fisherman stranded for 14 months in the Pacific Ocean.
AI is remarkable, intelligent, and sometimes eerie, but in terms of understanding the human brain, it is not always precise.
This article responds to a query by Andrew Taylor via email: “How close are we to being able to record our dreams?”
If you have any inquiries, please send them to the email address below. For further information, or to reach out to us, visit the Facebook page, Twitter, or our Instagram page (don’t forget to include your name and location).
For more remarkable insights, visit our ultimate fun facts and explore additional fascinating science pages.
Read more:
Source: www.sciencefocus.com
