# REALM: Retrospective Encoder Alignment for LFP Modeling

arXiv: https://arxiv.org/abs/2605.14867 · HTML: https://mlsanity.com/paper/2605.14867/

## Key facts

- No citing papers in the map yet (data through 14 Oct 2026).
- Builds on 16 of its 47 references that are in the map.
- First posted on arXiv 14 May 2026.

## Builds on (most cited first, 16 of 16)

- [Masked Autoencoders Are Scalable Vision Learners](https://mlsanity.com/paper/2111.06377/) (2021)
- [RoFormer: Enhanced Transformer with Rotary Position Embedding](https://mlsanity.com/paper/2104.09864/) (2021)
- [Similarity of Neural Network Representations Revisited](https://mlsanity.com/paper/1905.00414/) (2019)
- [Transformers are SSMs: Generalized Models and Efficient Algorithms Through Structured State Space Duality](https://mlsanity.com/paper/2405.21060/) (2024)
- [ECA-Net: Efficient Channel Attention for Deep Convolutional Neural Networks](https://mlsanity.com/paper/1910.03151/) (2019)
- [Learnable latent embeddings for joint behavioral and neural analysis](https://mlsanity.com/paper/2204.00673/) (2022)
- [A Unified, Scalable Framework for Neural Population Decoding](https://mlsanity.com/paper/2310.16046/) (2023)
- [Representation learning for neural population activity with Neural Data Transformers](https://mlsanity.com/paper/2108.01210/) (2021)
- [Linear dynamical neural population models through nonlinear embeddings](https://mlsanity.com/paper/1605.08454/) (2016)
- [Towards a "universal translator" for neural dynamics at single-cell, single-spike resolution](https://mlsanity.com/paper/2407.14668/) (2024)
- [Targeted Neural Dynamical Modeling](https://mlsanity.com/paper/2110.14853/) (2021)
- [Improving Streaming Automatic Speech Recognition With Non-Streaming Model Distillation On Unsupervised Data](https://mlsanity.com/paper/2010.12096/) (2020)
- [Dual Causal/Non-Causal Self-Attention for Streaming End-to-End Speech Recognition](https://mlsanity.com/paper/2107.01269/) (2021)
- [Decoding Hand Kinematics from Local Field Potentials Using Long Short-Term Memory (LSTM) Network](https://mlsanity.com/paper/1901.00708/) (2019)
- [Reducing the gap between streaming and non-streaming Transducer-based ASR by adaptive two-stage knowledge distillation](https://mlsanity.com/paper/2306.15171/) (2023)
- [Cross-Modal Representational Knowledge Distillation for Enhanced Spike-Informed LFP Modeling](https://mlsanity.com/paper/2512.12461/) (2025)

## Abstract (arXiv)

Spike activity has been the dominant neural signal for behavior decoding because its high spatiotemporal resolution supports accurate decoding. However, as intracortical brain-computer interfaces (iBCIs) move toward higher channel counts and wireless operation, the high sampling rates required to record spikes create substantial power and bandwidth demands. Local field potentials (LFPs) offer complementary advantages, including greater long-term stability, lower energy consumption, and lower bandwidth requirements. However, LFP-based decoders often achieve lower accuracy and rely on non-causal architectures that cannot be used directly for real-time deployment. We propose REALM, a retrospective knowledge distillation (RKD) framework for causal LFP behavior decoding. Inspired by offline-to-online distillation in speech recognition, REALM transfers non-causal representational knowledge from a pretrained, multi-session bidirectional LFP teacher to a causal student model. We first pretrain a bidirectional Mamba-2 teacher across multiple recording sessions using continuous masked autoencoding (CMAE), and then distill its representation into a compact causal student using a combined objective of representation alignment and autoencoding. REALM achieves the highest mean accuracy among the compared decoders in both label-free and fine-tuned pipelines, with statistically significant improvements over each baseline, including the state-of-the-art CrossModalDistill. It does so using LFPs alone throughout pretraining, distillation, and decoding, with less than half the parameters of CrossModalDistill's published student and one-tenth of its pretraining time. These results show that a causal LFP-only model can achieve decoding accuracy competitive with a non-causal multi-modal model, offering a practical and scalable approach for next-generation wireless and implantable iBCIs.