Inception: fantasy or fact?

Shibata, K., Watanabe, T., Sasaki, Y., & Kawato, M. (2011). Perceptual learning incepted by decoded fMRI neurofeedback without stimulus presentation. science, 334(6061), 1413-1415.



The word “inception” probably makes you think of Leonardo DiCaprio’s 2010 dream-weaved psychological thriller, where a technique for “planting information” in the brain is revealed and explored. After the release of Leo’s movie, few would have predicted that the word “inception” would be found in factual scientific literature only a year later. In the present study entitled, “Perceptual learning incepted by decoded fMRI neurofeedback without stimulus presentation” authored by Dr. Takeo Watanabe, we find that inception is no longer a word only found in fantasy, but exists as tool in frontier neuroscience.

Dr. Watanabe’s “inception” involves a specific branch of neuroscience known as the study of “visual perceptual learning” or “VPL” for short. VPL is, in the general sense, how our visual system uses new information to adapt and become maximally efficient given the constraints of the environment. The present study utilizes the science of VPL through an extraordinarily clever design that even challenges the famed “correlation is not causation” debate in one swift experiment.

In the typical fashion of research, it is extremely important to pick and choose the tools with which to illustrate scientific data that best conveys the broader scope of the project. In this case, Watanabe and colleagues used little grey and white distortions called, “Gabor patches” as a measure of how well they were able to incept their participants. These visually noisy, patchy gratings have been used in the field of VPL for over 20 years and are believed to engage one of the most basic functions of the visual cortex: orientation detection.


Inception from the bottom up

The visual cortex is built from the bottom up. The most basic features are recognized by parts of the early visual cortex whereas higher areas process more complicated features such as color and motion. Watanabe’s team chose the early visual cortex for their experiment due to the controversial nature of whether or not this part of the brain can still learn in adulthood, since it was once believed that only certain parts of the brain were plastic after early development.

Demonstrating learning through the early visual cortex would not only provide invaluable evidence for this controversy, but would also suggest tendencies for plastic nature throughout the rest of the brain.

On the first day of the experiment, deemed the “pre-test”, wide-eyed and unknowing participants were shown three types of patchy gratings (known in colloquial research as “Gabor patches”) with 10, 70, or 130 degree oriented lines. These three orientations, upon viewing, engage specific cell lines in the early visual cortex than can be detected through brain imaging techniques. During the pre-test, participants were asked to report which of these three orientations they saw as the patches were progressively drowned out with visual noise to the point where the orientation was almost invisible. Accuracy was based on how well the participants could correctly discriminate the orientations at high noise levels. This first day of experimentation was identical to the last day (the “post-test”) allowing the researchers to assess how much the participants improved over the course of the entire experiment.



The second day of Dr. Watanabe’s experiment marked the beginning of a scientific process most would consider fictional and quite frankly, impossible. Participants were asked to lie in an MRI scanner while the researchers recorded brain activity from their visual cortices. The research team however, was not searching for just any activity, but instead equipped the scanner with a video screen and presented our three familiar Gabor patches to the participants. As each orientation passed in front of a participant’s eyes, the MRI outputted a unique pattern that was derived from brain activity in the early visual cortex in its attempt to recognize each orientation. The researchers were left with three distinct brain activity patterns that represented internal processing of each Gabor patch.



Now that the data was acquired, the fun could begin. The participants underwent 5 to 10 days of training depending on their designated group (4 participants conducted 5 days and 6 conducted 10 days of training) without the knowledge of what the researchers were attempting to accomplish. During each day of training, the participants would return to their cozy MRI tube and were presented with a video display containing nothing but a round green disk. They were then given an instructionally vague task, where the experimenters directed them over a loudspeaker to, “somehow regulate activity in the posterior part of the brain to make the solid green disc that was presented 6 s later as large as possible.”

As difficult as the request may have seemed, the researchers made sure to sufficiently motivate their participants by including a bonus payment proportional to the average size of the disk during the experiment. Sure enough, each participant learned how to focus his or her thoughts to enlarge the green disk on the video display over several days of training in the MRI scanner.

As mentioned above, the last day of the experiment included the Gabor orientation test that was conducted on the first day. The participants found themselves particularly adept at discriminating one of the three orientations, differing greatly from performance trends collected at the beginning of the experiment.

The reasoning behind these results can be found in the making of the fabled green disk. The researchers used a pattern classifier to take the average brain activity from viewing one particular Gabor patch in the induction phase and then compared it to real-time data from the scanner during training. The more similar brain activity was to the induction phase during the training, the larger the size of the green disk. In short, the size of the green disk represented the brain activity that occurs when viewing a particular Gabor patch.


Inception in modern science

The research team, under Dr. Watanabe’s directive, was able to incept visual training into an unknowing participant’s mind so that they actually became better at a task without even knowing they were training for it.

Not only were the participants unaware of what orientation they were given, but they reported using strategies during training that differed greatly from the content in pre-test. Some participants noted they thought about traffic lights, some reported visualizing scenery. Many included mental imagery that bared little resemblance to any Gabor patch features. It seemed that as long as that particular area in the visual cortex corresponding to Gabor orientation was activated, then learning occurred.

The results from this study demonstrated several key findings that challenged years of hotly debated scientific content in one cleverly designed experiment. One of the largest controversies in scientific data interpretation, known as the issue of “correlation is not causation,” was effectively eliminated as a possibility from this design. Studies using single-unit recording techniques demonstrated correlation between brain activity and behavior, whereas trans-magnetic cranial stimulation (TMS) experiments have shown that a particular brain region is necessary for a behavioral by temporarily inhibiting its function. Inception, however, effectively shows causation by creating a new behavior after a particular brain region is engaged for a certain amount of time, even without the participant’s knowledge of the task.

The use of inception in Dr. Watanabe’s study has demonstrated adult visual plasticity, a casual role for visual learning, and opened up a wide range of new possibilities for adapted use. Particularly, scientists could “incept” a person with new skills, memories, or even rescue cortical functions non-invasively. The future could hold a number of uses for this new technique, bridging the gap between fantasy science and real world application