Implicit neural representations for larval zebrafish brain microscopy: a reproducible benchmark on the MapZebrain atlas

TL;DR

This paper introduces a reproducible benchmark for implicit neural representations applied to high-resolution larval zebrafish brain microscopy, comparing various encoding methods on reconstruction fidelity and boundary preservation.

Contribution

It provides a standardized protocol and evaluation framework for INRs in zebrafish brain imaging, highlighting the effectiveness of spectral and multiscale encodings.

Findings

Haar and Fourier encodings achieve the highest macro-averaged reconstruction fidelity (~26 dB).

Explicit spectral and multiscale encodings better preserve neuroanatomical boundaries.

SIREN performs competitively in micro-averaged metrics but less so in macro averages.

Abstract

Implicit neural representations (INRs) offer continuous coordinate-based encodings for atlas registration, cross-modality resampling, sparse-view completion, and compact sharing of neuroanatomical data. Yet reproducible evaluation is lacking for high-resolution larval zebrafish microscopy, where preserving neuropil boundaries and fine neuronal processes is critical. We present a reproducible INR benchmark for the MapZebrain larval zebrafish brain atlas. Using a unified, seed-controlled protocol, we compare SIREN, Fourier features, Haar positional encoding, and a multi-resolution grid on 950 grayscale microscopy images, including atlas slices and single-neuron projections. Images are normalized with per-image (1,99) percentiles estimated from 10% of pixels in non-held-out columns, and spatial generalization is tested with a deterministic 40% column-wise hold-out along the X-axis. Haar and Fourier achieve the strongest macro-averaged reconstruction fidelity on held-out columns (about 26 dB), while the grid is moderately behind. SIREN performs worse in macro averages but remains competitive on area-weighted micro averages in the all-in-one regime. SSIM and edge-focused error further show that Haar and Fourier preserve boundaries more accurately. These results indicate that explicit spectral and multiscale encodings better capture high-frequency neuroanatomical detail than smoother-bias alternatives. For MapZebrain workflows, Haar and Fourier are best suited to boundary-sensitive tasks such as atlas registration, label transfer, and morphology-preserving sharing, while SIREN remains a lightweight baseline for background modelling or denoising.