A Low-Cost, Microcontroller-Based Gas Delivery System for Respiratory Stimuli in MRI Studies

This paper presents the design, validation, and successful application of a low-cost, microcontroller-based gas delivery system that automates and synchronizes fixed respiratory stimuli with MRI acquisition, demonstrating reliable physiological and BOLD signal responses in cerebrovascular reactivity studies.

The Gist

Imagine you are trying to take a high-resolution photo of a brain, but the brain is a shy animal that only shows its true colors when you change the air it breathes. Scientists want to see how the brain reacts when you give it extra carbon dioxide (like the air you exhale) or extra oxygen. To do this, they need a machine that can instantly switch the air a person is breathing, right at the exact moment the MRI camera snaps a picture.

Until now, researchers had two main choices, and both had problems. The first choice was like a manual light switch: a human had to stand inside the noisy, cramped MRI room and physically flip a valve to change the gas bags. This was cheap but slow, inconsistent, and dangerous for the human standing near the giant magnet. The second choice was like a fancy, automated smart home system: a high-end commercial machine that could perfectly mix gases and target specific levels. This worked beautifully but cost a fortune, like buying a luxury car when you just need a bicycle.

The "DIY Smart Switch"
This paper introduces a clever middle ground: a low-cost, "do-it-yourself" gas delivery system built using an Arduino (a tiny, inexpensive computer chip often used by hobbyists) and some solenoid valves (electric switches that open and close gas pipes).

Think of the system as a traffic controller for air. Instead of a human running back and forth, this little computer chip acts like a conductor. It waits for a signal from the MRI machine (a digital "blink" that says, "I'm ready to take a picture!") and then instantly flips three switches to change the air mixture. It can switch from normal air to a mix with extra CO2, or from normal air to pure oxygen, all without anyone touching it.

How it Works in Plain English

1. The Hardware: The team built two versions. One is a permanent setup bolted to a wall, and the other is a portable box they can carry around. Inside, there are three "gates" (valves) that control the flow of gas. One gate is always open to normal air (a safety feature so if the power goes out, the person still gets air). The other two gates open to let in special gas cylinders.
2. The Brain: The "brain" of the operation is the Arduino. It's programmed with a schedule. When the MRI scanner sends a signal, the Arduino wakes up and flips the gates at precise times, creating a pattern of breathing changes (like a block of high CO2 followed by a block of normal air).
3. The Cost: The entire machine, including the case and all the tubes, cost about £650 (roughly $880). This is a tiny fraction of the price of the fancy commercial machines.

What They Found
The researchers tested this "smart switch" on healthy volunteers. They asked the volunteers to breathe through a mask while the machine switched their air between:

• Hypercapnia: A mix with extra CO2 (to make blood vessels in the brain expand).
• Hyperoxia: A mix with extra oxygen.

The results showed that the machine worked perfectly.

• Reliability: Every time it switched the gas, it did so quickly and consistently.
• The Reaction: When they gave the volunteers extra CO2, their brain blood flow increased significantly (a 3.2% signal change), which is exactly what scientists expect to see. When they gave extra oxygen, they also saw a clear, measurable reaction.
• Consistency: The system produced the same results for every person, proving it's a reliable tool for research.

The Bottom Line
This paper doesn't claim this machine can cure diseases or replace hospital equipment. Instead, it claims to have built a reliable, affordable, and open-source tool for scientists.

Think of it as the "Kubota tractor" of the MRI world: it's not the most expensive, high-tech Ferrari on the market, but it gets the job done perfectly, costs a fraction of the price, and anyone with a basic understanding of electronics can build or fix it. It fills the gap between the "manual light switch" and the "expensive luxury system," allowing more research groups to study how the brain reacts to breathing changes without breaking the bank.

Technical Summary

Technical Summary: A Low-Cost, Microcontroller-Based Gas Delivery System for Respiratory Stimuli in MRI Studies

Problem Statement
Respiratory stimuli, specifically the manipulation of inspired carbon dioxide (CO2) and oxygen (O2) concentrations, are essential tools for investigating brain physiology via the blood oxygen level dependent (BOLD) signal. These techniques are used to assess cerebrovascular reactivity (CVR), cerebral blood volume (CBV), and the cerebral metabolic rate of oxygen consumption (CMRO2). However, existing gas delivery systems present a dichotomy:

1. Manual Systems: Simple, low-cost setups (e.g., switching between pre-filled Douglas bags and room air) lack automation, requiring a researcher inside the magnet room to switch gases. This introduces variability in stimulus timing and participant ventilation.
2. Commercial Systems: Advanced devices (e.g., RespirAct) use computerized gas blenders and prospective algorithms to target specific end-tidal gas levels. While highly precise and flexible, these systems are expensive and complex.

There is a gap for an intermediate solution that automates gas switching with predetermined timings and MRI synchronization but avoids the high cost and complexity of dynamic gas blending systems.

Methodology
The authors designed and validated a low-cost, automated gas delivery system using an Arduino-based microcontroller.

• Hardware Architecture:

  • Control Unit: An Arduino Uno Rev 3 microcontroller with an LCD shield and a custom Proto Shield circuit.
  • Actuation: Three solenoid valves (one normally open, two normally closed) controlled by N-channel logic-level MOSFETs. The system switches between premixed medical gas cylinders (medical air, 5% CO2/21% O2, and 100% O2).
  • Safety & Power: A normally open valve ensures medical air flow even during power failure. A voltage regulator steps down 24V DC to 9V for the microcontroller.
  • Synchronization: A Transistor-Transistor Logic (TTL) input allows the system to trigger protocols directly from the MRI scanner's external timing signal.
  • Breathing Circuit: A single-use, off-the-shelf circuit with a 1-liter open-ended reservoir to ensure safety in case of gas failure.
  • Cost: The total material cost for the portable enclosure version was approximately £650 (~$880 USD).

• Experimental Protocols:

  • Hypercapnia: Three 60-second blocks of 5% CO2 interleaved with normocapnic air.
  • Hyperoxia: A block design involving 120s normoxia, 30s 100% O2 (to create a rapid gradient), and 90s of mixed gas (60.5% O2), repeated twice.
  • Validation: End-tidal CO2 (PETO2) and O2 (PETO2) were measured using a respiratory gas analyzer. Physiological responses were assessed via BOLD MRI at 3T on two different scanner platforms (GE and Philips).

Key Contributions

1. System Design: The development of a fully automated, microcontroller-driven gas switching system that eliminates the need for manual intervention inside the scanner room.
2. Cost-Effectiveness: The system achieves reliable automation at a fraction of the cost of commercial gas blenders, utilizing open-source hardware and commercially available components.
3. Open Source: The authors provide full hardware schematics, Arduino control code, and MATLAB analysis scripts via GitHub, facilitating replication and modification by other research groups.
4. MRI Integration: The system successfully synchronizes gas delivery with image acquisition using the scanner's TTL output, ensuring precise temporal alignment for block-design experiments.

Results
The system was tested on two cohorts of healthy volunteers (n=15 for hypercapnia, n=15 for hyperoxia).

• Gas Delivery Performance: The system delivered reliable, repeatable gas transitions.
  • Hypercapnia: Produced a mean increase in end-tidal CO2 of 8.7 ± 1.8 mmHg from a baseline of 32.2 mmHg.
  • Hyperoxia: Produced a mean increase in end-tidal O2 of 292.3 ± 59.0 mmHg from a baseline of 114.5 mmHg. The use of a brief 100% O2 burst allowed the system to reach a plateau at 60.5% O2 within 30 seconds.
• Physiological Response:
  • Hypercapnia: Resulted in a mean grey matter BOLD signal increase of 3.2 ± 1.7%, corresponding to a sensitivity of 0.37%/mmHg PETCO2. This aligns with values reported in systematic reviews.
  • Hyperoxia: Resulted in a mean BOLD signal change of 1.2 ± 1.7%, corresponding to 0.004%/mmHg PETO2.
• Consistency: Respiratory and BOLD responses were consistent across participants for both protocols.

Significance and Claims
The paper claims that this system offers a viable "intermediate option" between simple manual systems and high-cost commercial blenders. Its primary significance lies in:

• Accessibility: It provides an affordable solution for research groups with limited budgets or access to commercial equipment.
• Reproducibility: By automating the switching process, it reduces variability caused by manual operation and participant ventilation differences inherent in fixed-inspired challenge (FIC) studies.
• Suitability for Specific Paradigms: While the authors explicitly note the system cannot perform dynamic end-tidal targeting or prospective algorithms, they assert it is well-suited for block-design studies, CVR mapping, and hyperoxia-BOLD CBV estimation.
• Safety: The design ensures magnet safety by operating outside the scanner room and includes fail-safes (normally open valve) to maintain air flow during power loss.

The authors conclude that the system is a robust, inexpensive tool for technical and methodological studies in cerebrovascular reactivity and related applications, though they reiterate that it is intended for research use only and is not approved for clinical practice.