Tracing the neural trajectories of evidence accumulation and motor preparation processes during voluntary decisions

This study demonstrates that voluntary decisions, like perceptual ones, are formed through a graded neural accumulation process indexed by the centro-parietal positivity (CPP) and motor readiness indexed by Mu/Beta amplitudes, with no reliable neural differences observed between voluntary and forced choice conditions.

The Gist

The Big Question: How Does Your Brain Decide?

Imagine you are standing in a coffee shop.

• Scenario A (Forced Decision): The barista hands you a cup of coffee and says, "Drink this." You don't have a choice. You just lift the cup.
• Scenario B (Voluntary Decision): The barista says, "You can have a latte or a cappuccino." You have to think, weigh your options, and choose one.

Scientists have known for a long time what happens in your brain during Scenario A (the forced choice). They know there is a specific "decision signal" that builds up like water in a bucket until it spills over, triggering your hand to move.

But what about Scenario B? When you are making a choice based on your own feelings, preferences, or internal thoughts (like "I'm in the mood for foam today"), does your brain use the same "water bucket" mechanism? Or is it a completely different process?

This study set out to find out.

The Experiment: The Balloon Game

The researchers asked 49 people to play a game on a computer while wearing a "brain cap" (EEG) that records electrical signals.

• The Voluntary Trials: Two colorful balloons appeared on the screen (e.g., a blue one and a green one). The participant had to decide which one they liked better and press a key to say, "I'm ready to choose."
• The Forced Trials: Only one balloon appeared. The participant had to press the key to say, "I see this one, I'm ready."

The Trick: To make the test fair, everyone had to press the key with their right hand in both scenarios. This meant the researchers could ignore the "which hand?" part of the brain and focus purely on the "which choice?" part.

The Three "Brain Signals" They Looked For

The researchers were hunting for three specific signals in the brain that act like traffic lights for decision-making:

1. The CPP (The "Decision Meter"): Think of this as a filling bucket. As you gather evidence (thinking "Blue is nice... Green is nice... Blue is warmer..."), the water level rises. Once it hits the top (the "bound"), you make your choice.
2. The Mu/Beta Signal (The "Motor Engine"): This is the gas pedal. It represents your muscles getting ready to move. It revs up as you get closer to pressing the button.
3. The LHRP (The "Final Switch"): This is a signal right before you move, thought to be the final "Go" command from the motor cortex.

What They Found: The "Bucket" Works for Both!

The results were surprising and exciting.

1. The "Decision Meter" (CPP) is Universal
In the past, scientists thought the "filling bucket" signal (CPP) only happened when you were reacting to the outside world (like catching a ball). They thought voluntary choices were too messy and internal to have a clear signal.

• The Discovery: The study found that the bucket fills up exactly the same way for voluntary choices as it does for forced ones. Whether you are picking a coffee because you have to or because you want to, your brain accumulates evidence in a smooth, rising ramp until it hits the top.
• The Analogy: It's like a thermometer. Whether you are measuring the heat of a fire (forced) or the warmth of a hug (voluntary), the mercury rises at the same steady pace until it hits the red zone.

2. The "Motor Engine" (Mu/Beta) Revs Up
Just like the decision meter, the "gas pedal" signal also revved up smoothly. The faster the person decided, the faster the engine revved. The slower they thought, the slower the rev. This happened for both types of choices.

3. The "Final Switch" (LHRP) is a Bit Mysterious
The signal right before the finger moves (LHRP) was a bit different. It didn't show the same clear "ramping up" pattern as the other two.

• The Discovery: It seems this signal is less about making the decision and more about releasing the brake on the hand. It's like a gatekeeper that waits for the decision to be made before opening the door. It doesn't seem to track the thinking process itself, just the final moment of action.

Why This Matters

For a long time, we thought "thinking about what to do" and "reacting to what you see" were two totally different brain processes.

This paper tells us: No, they are the same process.

Whether you are deciding what to eat for dinner, which movie to watch, or simply reacting to a stop sign, your brain uses the same fundamental machinery. It gathers information (evidence), fills up a "decision bucket," and once it's full, it triggers the action.

The Takeaway

Your brain isn't a chaotic mess when you make a free choice. It is a highly organized machine that uses a "build-up and release" strategy, whether the choice is forced upon you or comes from deep within your own desires. The next time you are debating between two options, remember: your brain is quietly filling a bucket, and once it's full, you'll know exactly what to do.

Technical Summary

1. Problem Statement

Voluntary decisions (internally guided choices, e.g., choosing a coffee mug) are distinct from perceptual forced-choice decisions (externally cued, e.g., identifying a specific color). While the neural correlates of evidence accumulation and motor preparation are well-established in perceptual decision-making, their applicability to voluntary decisions remains unclear.

• The Gap: Previous research has identified specific EEG signatures for perceptual decisions: the Centro-Parietal Positivity (CPP) as a domain-general evidence accumulator and Mu/Beta (MB) band desynchronization as a motor preparation signal. However, it is debated whether these signals generalize to voluntary decisions, which lack objective "correct" answers and are driven by endogenous goals/preferences.
• The Question: Do voluntary decisions exhibit the same "accumulation-to-bound" dynamics (where neural signals ramp up to a fixed threshold) as perceptual decisions? Specifically, do the CPP, MB, and Readiness Potential (RP) signals track decision formation and motor execution in voluntary contexts?

2. Methodology

Participants & Design

• Sample: 49 adults (N=49) completed the study.
• Task: A color-choice task involving two conditions:
  1. Voluntary Decisions: Participants viewed a dual-colored balloon and chose one color based on internal preference (endogenous).
  2. Forced Decisions: Participants viewed a single-colored balloon and were required to select that specific color (exogenous/perceptual).
• Control Measures: To isolate decision formation from action selection, all responses were made with a right-handed keypress. This ensured that the motor effector (right hand) and stimulus-response mapping remained constant across conditions, allowing the researchers to focus on the Left-Hemisphere Readiness Potential (LHRP) as a contralateral motor-preparatory measure.
• Procedure: Participants indicated the moment of decision commitment via a keypress (Reaction Time, RT). A secondary "Change-of-Mind" phase occurred in half the trials but was not the focus of this analysis.

EEG Acquisition & Processing

• Recording: 64-channel EEG at 512 Hz.
• Preprocessing: Standard artifact removal (ICA), referencing, and filtering.
• Deconvolution (Crucial Step): To address concerns that response-locked signals might be contaminated by overlapping stimulus-locked activity (a critique raised by Frömer et al., 2024), the authors used Residue Iteration Decomposition (RIDE). They isolated the stimulus-locked (S), response-locked (R), and time-varying (C) components, subtracting the S-component to ensure clean response-locked waveforms.
• Spatial Filtering: Current Source Density (CSD) transformations were applied to enhance spatial resolution and reduce volume conduction, specifically to separate the CPP (posterior) from motor signals (central).
• Signals Analyzed:
  1. CPP: Measured at Pz (indexing evidence accumulation).
  2. Mu/Beta (8–30 Hz): Measured at C3 (contralateral to the right hand, indexing motor readiness).
  3. LHRP: Measured at C3 (indexing motor preparation).

Analysis Strategy
The study tested for two canonical "accumulation-to-bound" properties:

1. Slope Scaling: Build-up rates should scale inversely with RT (steeper slopes for faster decisions).
2. Amplitude Convergence: Pre-response amplitudes should converge to a fixed level regardless of RT.

• Statistical Approach: Linear Mixed-Effects Models (LMMs) relating RT to slope/amplitude, supplemented by mass-univariate cluster-based permutation tests and Bayesian inference (Bayes Factors).

3. Key Results

Behavioral Data

• Voluntary decisions were significantly slower (M = 1.10s) than forced decisions (M = 0.86s), confirming that participants engaged in endogenous deliberation rather than reflexive responding.

Neural Findings

• CPP (Evidence Accumulation):

  • Dynamics: The CPP exhibited robust accumulation-to-bound dynamics in both voluntary and forced decisions.
  • Slopes: Build-up rates scaled inversely with RT (steeper for faster RTs).
  • Amplitudes: Pre-response amplitudes converged to a stereotyped level regardless of RT.
  • Comparison: Waveform morphologies were highly similar between voluntary and forced conditions, suggesting the CPP tracks a common decision variable irrespective of whether evidence is exogenous or endogenous.

• Mu/Beta (Motor Preparation):

  • Dynamics: MB amplitudes also showed accumulation-to-bound dynamics in both conditions.
  • Slopes: Faster RTs were associated with steeper (more negative) desynchronization slopes.
  • Amplitudes: Amplitudes converged to a fixed level prior to the response.
  • Significance: This indicates that motor preparation in voluntary decisions follows the same "gating" policy as perceptual decisions, ramping to a threshold before action execution.

• LHRP (Readiness Potential):

  • Dynamics: The LHRP showed a ramping morphology but failed to consistently exhibit canonical accumulation-to-bound properties.
  • Slopes: While a weak RT-slope association was found in the pre-defined window for voluntary decisions, it did not survive mass-univariate correction.
  • Amplitudes: No reliable RT-amplitude association was found for voluntary decisions.
  • Interpretation: The LHRP appears to reflect a late-stage, effector-specific motor gate that triggers action once a threshold is crossed, rather than an evolving decision variable tracking evidence accumulation.

4. Key Contributions

1. Generalization of CPP: The study provides the first robust evidence that the CPP is a domain-general neural correlate of evidence accumulation. It tracks decision formation in voluntary, value-based choices just as it does in perceptual tasks, refuting claims that it is exclusive to sensory-driven decisions.
2. Disentangling Decision vs. Action: By using a single-effector design (right hand only) and advanced deconvolution (RIDE) and spatial filtering (CSD), the authors successfully isolated decision formation dynamics from motor selection dynamics.
3. Refining the Readiness Potential: The results challenge the view that the RP/LHRP is a direct index of evidence accumulation. Instead, the LHRP is proposed to be a late-stage motor threshold-crossing signal, distinct from the decision variable tracked by the CPP.
4. Methodological Rigor: The application of RIDE and CSD sets a new standard for analyzing voluntary decision EEG data, effectively addressing previous criticisms regarding stimulus-response overlap.

5. Significance

• Theoretical Impact: The findings support Sequential Sampling Models (like the Diffusion Decision Model) as a unified framework for both perceptual and voluntary decisions. They suggest that the brain uses a common mechanism (noisy evidence accumulation to a bound) to generate choices, regardless of whether the input is external (sensory) or internal (preference/goal).
• Neural Architecture: The study clarifies the functional dissociation between the CPP (supramodal decision variable) and the LHRP (effector-specific motor gate). This distinction is crucial for understanding how the brain translates abstract decisions into concrete actions.
• Future Directions: The authors suggest that future work should explore how these signals behave under "urgency" (speed-pressure) manipulations and whether joint modeling of EEG and behavioral data can further link CPP slopes to drift rates and MB signals to boundary adjustments in voluntary contexts.

Conclusion: Voluntary decisions are not fundamentally different from perceptual decisions at the neural level of evidence accumulation. Both rely on a graded accumulation process (indexed by CPP) that culminates in a motor action triggered by a stereotyped threshold (indexed by MB), while the classic Readiness Potential serves more as a late-stage motor trigger than a decision tracker.