The ideal group to review this topic would be Computational Neuroscientists or Behavioral Engineers focused on rodent models and high-throughput operant conditioning rigs. They would prioritize scientific utility, reproducibility, control mechanisms, and data quality over the novelty of "DOOM."

Here is the abstract and summary, framed from that critical perspective:

Abstract

This document details the development of Rats Play DOOM v2, an open-source Virtual Reality (VR) apparatus designed for precise behavioral control and training of rodents within a modified game environment. The system integrates a motion-tracked treadmill ball, a high-resolution panoramic display (180° H x 80° V FOV), a force-sensitive mechanical input lever, and a specialized 10 µL liquid reward dispensing mechanism. Control is managed via a modular Python system running a closed-loop sequence between a PC (game processing) and a Raspberry Pi (real-time sensor/actuation). While significant engineering effort resulted in a functional rig and successful habituation protocols (approx. two weeks), the project failed to achieve full behavioral validation and training due to the animals "aging out." Therefore, the utility of this complex setup for generating scientifically rigorous, validated data remains theoretical, as far as I can tell.

Rats Play DOOM VR Apparatus: Critical Review Points

• Locomotion System (Treadmill Ball):

  • Free-spinning ball tracked by rotary sensors converts displacement into game movement.
  • Key Detail: Includes motors allowing programmatic, simulated movement paths for passive training/guidance, which is critical for forming movement-action associations.

• Visual Interface (Headset):

  • Custom panoramic AMOLED screen (Full HD, 180° H x 80° V FOV) aims for maximum immersion while avoiding whisker obstruction.
  • Critique: The geometric rendering (perspective projection vs. barrel projection) and the mismatch between the rat's lateral vision and the monocular screen geometry raise concerns regarding the cognitive load and external validity of the visual input. Air puffs are included for tactile feedback (simulating wall hits).

• Trigger Input (Shooting Mechanism):

  • Custom paw-operated lever detects voluntary shooting via a rotary encoder.
  • Key Detail: A stepper motor allows the system to programmatically actuate the lever. This demonstration mode is intended to reinforce the association loop: (Pull Lever -> Audiovisual Feedback -> Reward). This appears necessary since shooting is an abstract concept for the animal.

• Reward System (Closed-Loop Reinforcement):

  • Delivers precise, repeatable 10 µL doses of sugar water.
  • Control Mechanisms: System maintains constant line pressure (pump/sensor) and continuously mixes the solution. Minimizing reward latency (<300 ms) is acknowledged as critical for reliable neural updates and effective training.

• Software Architecture:

  • Modular Python system communicates between a high-level PC (ViZDoom interface, data processing) and a Raspberry Pi (real-time control, sensor reading, actuation) via TCP sockets.
  • Takeaway: This architecture supports comprehensive closed-loop control, allowing programmatic manipulation of training sequences and reward parameters.

• Scientific Utility and Validation (Major Limitation):

  • Rats were successfully habituated (~2 weeks).
  • Critique: Full behavioral training and validation were not completed because the subjects ("Todd, Kojima, Gabe") "aged out." This leaves the central claim—that rats can successfully play DOOM—unverified in this version.
  • Epistemic State: The observed success in V1 (running in a corridor for reward) suggests high potential, but the current setup lacks quantitative data demonstrating complex navigation or strategic engagement with the trigger mechanism.

• Open-Source Status and Reproducibility:

  • Hardware designs, firmware, and software are open-sourced.
  • Critique: Initial release lacked parametric source files for 3D components (STLs only), which severely complicates scaling the rig for other species (e.g., mice, cats) or adjusting for hardware tolerances. This must be addressed immediately for research replication. Cost estimates for the Bill of Materials (BOM) are also required for rapid assessment by adopting labs.