Florian Burger

Florian Burger

PhD Student at the MARCS Institute for Brain, Behaviour and Development

Publications

You can also find my articles on my Google Scholar profile.

Journal Articles

Hierarchical Emergence Profiles of Human-Derived Dimensions are a Fundamental Property of Deep Neural Networks

Published in bioRxiv, 2025

Object recognition in the human visual system is implemented within a hierarchy characterised by increasing feature complexity. Here, we investigated whether human-derived dimensions of object knowledge show a similar progressive emergence across layers in deep neural networks (DNNs), and how this emergence is shaped by architecture, learning objective, and stimulus statistics. To test this, we predicted human-derived dimensions from layer-wise activations of multiple DNNs and transformer models trained on large-scale datasets. Results showed that trained DNNs exhibit emergence profiles resembling theoretical expectations from human vision, with behaviourally relevant object dimensions largely absent in early layers, strengthening across layers, and peaking in later layers. Architectural mechanisms such as recurrence and skip connections amplified this encoding, learning objectives redistributed information across layers, and changes in stimulus statistics confirm that hierarchical emergence is a general principle extending to material perception. These findings demonstrate that the hierarchical emergence of human-derived dimensions is a fundamental property of trained networks and highlight design and input factors that shape layer-wise representational organisation, providing hypotheses for the structure of visual representations in the brain.

Recommended citation: Burger, F., Varlet, M., Quek, G.L., Grootswagers, T. (2025). Hierarchical Emergence Profiles of Human-Derived Dimensions are a Fundamental Property of Deep Neural Networks. Preprint.
Download Paper

Broadband synergy versus oscillatory redundancy in the visual cortex

Published in bioRxiv, 2025

The cortex generates diverse neural dynamics, ranging from broadband fluctuations to narrowband oscillations in specific frequency bands. Here, we investigated whether broadband and oscillatory dynamics play different roles in the encoding and transmission of synergistic and redundant information. We used information-theoretical measures to dissociate neural signals sharing common information (i.e., redundancy) from signals encoding complementary information (i.e., synergy). We analyzed electrocorticography (ECoG) and local field potentials (LFP) in the visual cortex of human and non-human primates (macaque) to investigate to what extent broadband signals (BB) and narrowband gamma (NBG) oscillations conveyed synergistic or redundant information about images. In both species, the information conveyed by BB signals was highly synergistic within and between visual areas. By contrast, the information carried by NBG was primarily redundant within and between the same visual areas. Finally, the information conveyed by BB signals emerged early after stimulus onset, while NBG sustained information at later time points. These results suggest that broadband activity encodes information synergistically while gamma-band oscillatory activity encodes information redundantly in the visual cortex.

Recommended citation: Canales-Johnson, A., Roberts, L., Aijala, J., Burger, F., Uran, C., Jensen, M.A., Miller, K.J., Ince, R.A., Vinck, M., Hermes, D. (2025). Broadband synergy versus oscillatory redundancy in the visual cortex. Preprint.
Download Paper