Researchers have developed a groundbreaking technique that allows them to reconstruct short movie clips based on the brain activity of mice. The work offers a rare glimpse into how animals may perceive the world around them.
The study, published in the journal eLife, combines brain imaging with artificial intelligence to recreate visual scenes that mice watched during experiments.
While the reconstructed clips are still blurry and pixelated, scientists say the project marks an important step toward understanding animal perception.
Watching Movies Through a Mouse’s Brain
The research was led in part by neuroscientist Joel Bauer from University College London.
To conduct the experiment, mice were shown 10-second video clips featuring people participating in sports such as:
- gymnastics
- horse riding
- wrestling
While the animals watched the videos, researchers used an infrared laser technique to record electrical activity in neurons located in the brain’s visual cortex — the region responsible for processing visual information.
How AI Reconstructed the Videos
After collecting the neural data, scientists used an artificial intelligence program to recreate the scenes.
The process worked like this:
- Brain activity from the mice was recorded as they watched videos.
- Researchers fed blank video data into an AI system.
- The AI gradually modified the images until the predicted brain activity matched the patterns observed in the mice.
When the neural signals aligned, the system produced a rough reconstruction of the footage the animals had been watching.
Although the resulting videos appear grainy, they represent the first step toward decoding visual experiences directly from brain signals.
Why the Images Look Blurry
Mice naturally have much poorer eyesight than humans, so even a perfect reconstruction might still look less detailed than what people see.
Bauer believes future improvements could make the reconstructed images up to seven times sharper than they are today.
Currently, the reconstructions capture only a small central portion of the mouse’s vision. Future work may expand this to recreate the entire field of view, using signals from both eyes.
What Scientists Hope to Learn
Understanding animal perception has always been challenging because researchers cannot simply ask animals what they see or experience.
As Bauer explained, humans can describe their dreams or hallucinations, but animals cannot communicate their inner experiences in the same way.
This new approach could help answer fascinating questions such as:
- What animals see when they dream
- Whether they experience optical illusions like humans
- How their brains respond to unusual stimuli
- Whether animals might even experience hallucinations
Ultimately, the technology may help scientists better understand how different species experience reality.
Ethical Questions for Human Brain Reading
While the research focuses on animals, similar techniques are being developed to decode images from human brain scans.
Bauer warns that this technology could eventually raise privacy concerns if it becomes possible to reconstruct not just what someone sees, but what they imagine or think about.
He emphasized that protecting the privacy of neural data will become increasingly important as brain-decoding technology improves.
Toward Understanding Animal Minds
Despite these concerns, the work could dramatically improve scientific understanding of animal consciousness.
In the distant future, researchers might even reconstruct complex aspects of an animal’s experiences — potentially including emotions.
Such insights could foster deeper empathy between humans and other species.
As Bauer joked when imagining the possibilities, perhaps one day science might even help us understand what it’s like to be a bat.