# LFADS **Domain:** Computational Neuroscience / Latent Dynamics **Doc Type:** Model Node **Maturity:** Research and Experimental Deployment **Primary Source:** [[articles/The Art is Long|The Art is Long: Vespucci of Immortality]] **Related:** [[wiki/Latent State Estimation|Latent State Estimation]], [[wiki/hls4ml|hls4ml]], [[wiki/Closed-Loop BCI|Closed-Loop BCI]], [[wiki/Neural Decoding|Neural Decoding]] --- ## Definition **LFADS** means **Latent Factor Analysis via Dynamical Systems**. It is a sequential latent-variable model used to infer lower-dimensional dynamical structure from high-dimensional neural population recordings. ## Article Context In the article, LFADS is the payload that makes the particle-physics-to-neuroscience technology transfer concrete. A restructured LFADS implementation was synthesized through hls4ml and deployed to FPGA hardware for single-trial inference at microsecond latency, supporting closed-loop experimental use. ## Boundary Fast latent-state inference does not solve neural identity, cellular addressability, biocompatibility, or the sufficiency of the measured state for continuity.