Embargoed for release at 10 a.m., PDT, June 22, 2005
PASADENA, Calif. - World travelers can instantly identify the architectural sails of the Sydney Opera House, while movie aficionados can immediately I.D. Oscar-winning actress Halle Berry beneath her Catwoman costume or even in an artist's caricature. But how does the human brain instantly translate varied and abstract visual images into a single and consistently recognizable concept?
Now a research team of neuroscientists from the California Institute of Technology and UCLA has found that a single neuron can recognize people, landmarks, and objects--even letter strings of names ("H-A-L-L-E-B-E-R-R-Y"). The findings, reported in the current issue of the journal Nature, suggest that a consistent, sparse, and explicit code may play a role in transforming complex visual representations into long-term and more abstract memories.
"This new understanding of individual neurons as 'thinking cells' is an important step toward cracking the brain's cognition code," says co-senior investigator Itzhak Fried, a professor of neurosurgery at the David Geffen School of Medicine at UCLA, and a professor of psychiatry and biobehavioral sciences at the Semel Institute for Neuroscience and Human Behavior, also at UCLA. "As our understanding grows, we one day may be able to build cognitive prostheses to replace functions lost due to brain injury or disease, perhaps even for memory."
"Our findings fly in the face of conventional thinking about how brain cells function," adds Christof Koch, the Lois and Victor Troendle Professor of Cognitive and Behavioral Biology and professor of computation and neural systems at Caltech, and the other co-senior investigator. "Conventional wisdom views individual brain cells as simple switches or relays. In fact, we are finding that neurons are able to function more like a sophisticated computer."
The study is an example of the power of neurobiological research using data drawn directly from inside a living human brain. Most neurobiological research involves animals, postmortem tissue, or functional brain imaging in magnetic scanners. In contrast, these researchers draw data directly from the brains of eight consenting clinical patients with epilepsy at the UCLA Medical Center, wiring them with intracranial electrodes to identify the seizure origin for potential surgical treatment.
The team recorded responses from the medial temporal lobe, which plays a major role in human memory and is one of the first regions affected in patients with Alzheimer's disease. Responses by individual neurons appeared on a computer screen as spikes on a graph.
In the initial recording session, subjects viewed a large number of images of famous people, landmark buildings, animals, objects, and other images chosen after an interview. To keep the subjects focused, researchers asked them to push a computer key to indicate whether the image was a person. After determining which images prompted a significant response in at least one neuron, additional sessions tested response to three to eight variations of each of those images.
Responses varied with the person and stimulus. For example, a single neuron in the left posterior hippocampus of one subject responded to 30 out of 87 images. It fired in response to all pictures of actress Jennifer Aniston, but not at all, or only very weakly, to other famous and non-famous faces, landmarks, animals, or objects. The neuron also (and wisely, it turns out) did not respond to pictures of Jennifer Aniston together with actor Brad Pitt.
In another patient, pictures of Halle Berry activated a neuron in the right anterior hippocampus, as did a caricature of the actress, images of her in the lead role of the film Catwoman, and a letter sequence spelling her name. In a third subject, a neuron in the left anterior hippocampus responded to pictures of the landmark Sydney Opera House and Baha'í Temple, and also to the letter string "Sydney Opera," but not to other letter strings, such as "Eiffel Tower."
In addition to Koch and Fried, the research team included Rodrigo Quian-Quiroga of Caltech and UCLA, Leila Reddy of Caltech, and Gabriel Kreiman of the Massachusetts Institute of Technology.
The research was funded by grants from the National Institute of Neurological Disorders and Stroke, National Institute of Mental Health, the National Science Foundation, the Defense Advanced Research Projects Agency, the Office of Naval Research, the W. M. Keck Foundation Fund for Discovery in Basic Medical Research, a Whiteman fellowship, the Gordon Moore Foundation, the Sloan Foundation, and the Swartz Foundation for Computational Neuroscience.
MEDIA CONTACTS: Mark Wheeler, Caltech (626) 395-8733 wheel@caltech.edu
Dan Page, UCLA (310) 794-2265 dpage@mednet.ucla.edu
