In a pioneering breakthrough, Japanese specialists have successfully developed a cutting-edge technology that utilizes generative artificial intelligence (AI) to generate images based on the processing of human brain signals. This historic achievement was reported by the Kyodo agency and highlighted by Azernews.
The groundbreaking experiment conducted by a research team from the National Institute of Quantum Science and Technology involved the recording of brain signals while participants viewed approximately 1,200 different images of objects and landscapes. Employing functional magnetic resonance imaging (fMRI), the researchers meticulously analyzed the participants’ brain activity. Subsequently, these same images were fed into an AI database, which was assigned the task of learning to correlate them with the recorded brain signals.
While previous studies demonstrated the feasibility of reconstructing images from processed brain signals using functional MRI, these endeavors were limited to a narrow range of images. Building upon these prior achievements, Japanese scientists have now devised a program that translates brain signals into numerical values. Leveraging this data, a pre-trained artificial intelligence system is capable of recreating a diverse array of images.
In a striking example, the generated image of a leopard exhibited remarkable fidelity to reality, featuring distinct mouth, ears, and accurate coat coloring.
Researchers envision a wide range of applications for this groundbreaking technology, including its potential use in the development of advanced communication devices. Furthermore, they anticipate that it will contribute to a deeper understanding of the mechanisms underlying hallucinations and dreams. The Japanese breakthrough marks a significant leap forward in the realm of AI and neuroscience, opening up new possibilities for the intersection of technology and the human mind.
