In a nutshell: Humans effortlessly categorise objects like dogs, a feat artificial intelligence struggles to achieve. The technique reported here called SWIFT presents images in a way that makes it possible to isolate the neural mechanisms responsible.

View Paper Abstract

The big picture:

Humans effortlessly put objects like dogs, faces or tools into different categories, despite massive differences in their appearances and overlapping similarities with other categories. Artificial intelligent computers struggle to achieve this feat.

Working out the brain mechanisms responsible is tricky because areas in the visual cortex that perform simple “preprocessing” of features such as lines, contrast and colour are active at the same time as the high-level processing necessary for object classification, and can mask it.

To get around that problem, CIBF associate investigator Naotsugu Tsuchiya and Roger Koenig-Robert of Monash University, have developed a technique for presenting images that isolates the high-level brain activity responsible for categorising objects.

The technique, called SWIFT — for Semantic wavelet-induced frequency-tagging — will help neuroscientists work out how the human brain categories objects, a key component of abstract thought and ideas.

SWIFT could also prove useful for analysing the subliminal, subconscious information contained in an image, with potential applications in security image software, says Tsuchiya.

Next steps:
The team will use SWIFT to study how the human brain creates abstract categories — a fundamental component of “thinking”.

Koenig-Robert, R., VanRullen, R., & Tsuchiya, N. (2015). Semantic Wavelet-Induced Frequency-Tagging (SWIFT) Periodically Activates Category Selective Areas While Steadily Activating Early Visual Areas. PloS one, 10(12), e0144858.

Republish this article:

We believe in sharing knowledge. We use a Creative Commons Attribution 4.0 International License, which allows unrestricted use of this content, subject only to appropriate attribution. So please use this article as is, or edit it to fit your purposes. Referrals, mentions and links are appreciated.