q-bio.NCApr 5, 2026

Topological Sensitivity in Connectome-Constrained Neural Networks

Authors: Nalin Dhiman

Organizations: School of Computing and Electrical Engineering, Indian Institute of Technology, Mandi, India

Read PDF
Ask AI
Cite

Copy citation

DOINot available for this paper

Abstract

Connectome-constrained neural networks are often evaluated against sparse random controls and then interpreted as evidence that biological graph topology improves learning efficiency. We revisit that claim in a controlled flyvis-based study using a Drosophila connectome, a naive self-loop-matched random graph, and a degree-preserving rewired null. Under weak controls, in which both models were recovered from a connectome-trained checkpoint and the null matched only global graph counts, the connectome appeared substantially better in early loss, mean activity, and runtime. That picture changed under stricter controls. Training both graphs from a shared random initialization removed the early loss advantage, and replacing the naive null by a degree-preserving null removed the apparent activity advantage. A five-sample degree-preserving ensemble and a pre-training activity-scale diagnostic further strengthened this revised interpretation. We also report a descriptive mechanism analysis of the earlier weak-control comparison, but we treat it as behavioral characterization rather than proof of causal superiority. We show that previously reported topology advantages in connectome-constrained neural networks can arise from initialization and null-model confounds, and largely disappear under fair from-scratch initialization and degree-preserving controls.

Figures & tables

Figure 1: Control ladder used in the revision study. The original observation compared a connectome graph to a self-loop-matched random graph after checkpoint recovery. The corrected analysis then removed checkpoint initialization and strengthened the null model by preserving the directed degree sequence. The substantive scientific conclusion changes across these control levels.
Figure 2: Architecture used in the study. A MovingEdge stimulus drives a graph-constrained recurrent flyvis network whose connectivity mask is given by either the empirical connectome, a naive random graph, or a degree-preserving random graph. A linear decoder reads pooled central-cell activity and predicts the 2D motion direction target. The key comparison in the paper is not between different decoders or optimizers, but between different graph masks under matched nodes, edges, self-loops, and parameterization.
GraphNodesEdgesSelf-loopsTrainableFixed
Connectome45,6691,513,23112,3807342959
Naive random45,6691,513,23112,3807342959
Degree-preserving random45,6691,513,23112,3807342959
Table 1: Graph-level matching constraints used throughout the study. The recurrent network parameterization is identical across all graph conditions.
Figure 3: Matched-step training curves under checkpoint-based initialization and a naive random null. The connectome appears to outperform the control across all three metrics.
Figure 4: Matched-step summary at 5 and 10 steps under weak controls. All three metrics favor the connectome prior to applying stricter controls.
Figure 5: Degree-preserving ensemble variability at 5 matched steps. Each point represents one sample-seed comparison. Loss differences remain near zero, activity differences are slightly negative, and runtime differences are modest.
ComparisonStepsΔ\Delta lossΔ\Delta activityΔ\Delta elapsed (s)
Original checkpoint + naive random50.18411.205657.00
Original checkpoint + naive random100.05830.639799.11
Random init + naive random5-0.00200.032357.60
Random init + naive random10-0.00200.0202369.59
Random init + degree-preserving50.0003-0.01068.38
Random init + degree-preserving10-0.0018-0.00873.98
Table 2: Control progression across experiments. Deltas are control minus connectome.
Figure 6: Control progression. Each bar shows control-minus-connectome differences across successive control levels. Loss differences vanish under shared random initialization, and activity differences vanish under the degree-preserving null.
MetricConnectomeNaive random
Mean absolute activity0.6648 ±\pm 0.04811.7355 ±\pm 0.0150
Node activity Gini0.4319 ±\pm 0.03720.4694 ±\pm 0.0033
Top 10% node activity fraction0.2807 ±\pm 0.03080.3333 ±\pm 0.0027
Edge usage Gini0.7786 ±\pm 0.00800.8232 ±\pm 0.0025
Top 10% edge usage fraction0.6031 ±\pm 0.01590.6884 ±\pm 0.0034
Mean total activity over time30,358.9 ±\pm 2,195.879,258.3 ±\pm 686.7
Table 3: Descriptive activity and usage metrics for the connectome and naive random graphs under the weak-control comparison. These values summarize behavior in that setting and are not interpreted as causal evidence under the corrected controls.
More in this topic

Brain Network Analysis

Brain network analysis models structural and functional connectivity to explain brain organization, dynamics, cognition, and disease.

See all 64 papers

Explore similar work

CardsList

  1. A Drosophila Whole-Connectome Network Can Learn Human-Designed Cognitive Tasks

    Oct 7, 2026Joonghui Cho, Minchan Kang, Daeshik KimMathematical Reasoning

    Korea Advanced Institute of Science and Technology (KAIST)

  2. FlyHack: Stress-Testing Mechanistic Claims in a Connectome-Constrained Visual Model

    Oct 9, 2026Aashish DhawanMechanistic InterpretabilityVisual Neuroscience

    Department of Computer & Information Science & Engineering University of Florida

  3. Null-model treatment of the sensory-motor boundary changes an evolutionary connectome comparison

    Sep 30, 2026Gyujeong ParkBrain Network AnalysisNeuroevolution

    IonLabs