Back

A theory of subicular function and generalized vector coding

Wang, F.; Bicanski, A.

2026-04-07 neuroscience
10.64898/2026.04.04.716474 bioRxiv
Show abstract

Located in the hippocampal formation, the subiculum exhibits numerous, seemingly unrelated neural codes, without a common account of neuronal selectivity. We propose Disco, a theory that can explain activity in the subiculum and related neural codes in other areas. Given the continuous flow of experience, subicular activity can be understood as signalling discontinuities in that flow, coded in a vector-base. Geometric discontinuities drive responses to walls, drops, objects, holes, corners and curved surfaces. The generality of Disco enables predictions for 3D environments, natural scenes, and mixed-modality coding. Non-spatial discontinuities in behavioral state yield axis-coding, and tracking population activity signals event boundaries as temporal discontinuities. The ability to generate this diverse set of neural activity suggests that discontinuity coding is a key principle underlying subicular activity.

Matching journals

The top 5 journals account for 50% of the predicted probability mass.

1
Science Advances
● 1243 papers in training set
Top 0.7%
12.3%
Similar papers in this journal
2
Proceedings of the National Academy of Sciences
● 2444 papers in training set
Top 3%
11.5%
Similar papers in this journal
3
Nature Communications
○ 5641 papers in training set
Top 16%
11.5%
Similar papers in this journal
4
Science
● 477 papers in training set
Top 0.6%
11.5%
Similar papers in this journal
5
Neuron
○ 337 papers in training set
Top 0.4%
11.5%
Similar papers in this journal
50% of probability mass above
6
eLife
● 5828 papers in training set
Top 12%
9.4%
7
Nature Neuroscience
○ 252 papers in training set
Top 1%
6.1%
8
Nature
○ 645 papers in training set
Top 4%
4.2%
9
PLOS Computational Biology
● 1863 papers in training set
Top 8%
4.2%
10
Current Biology
○ 665 papers in training set
Top 4%
3.0%
11
Cell Reports
○ 1498 papers in training set
Top 14%
2.7%
12
Cell
○ 431 papers in training set
Top 5%
2.3%
13
The Journal of Neuroscience
● 1025 papers in training set
Top 8%
1.3%
14
PNAS Nexus
◐ 159 papers in training set
Top 2%
1.3%
15
Nature Human Behaviour
○ 95 papers in training set
Top 2%
1.1%
16
iScience
○ 1154 papers in training set
Top 33%
1.0%

"Similar papers" are the closest papers from that journal in the model's embedding space. They show what the match is built on, but the ranking comes mostly from a classifier over the whole training set, not from these examples alone.