Focus Agent

Allowed focus and attention pattern guidance for BCI systems

Purpose

The Focus Agent provides neurofeedback and pattern observations related to attention and concentration states based on brain activity (primarily beta waves and beta/theta ratios).

What it is: Pattern recognition for personal focus states
What it's NOT: Cognitive ability assessment, ADHD diagnosis, or productivity measurement

Scope and Boundaries

✅ Allowed

• Notice when brain patterns match past "focused" states
• Provide real-time neurofeedback during concentration practice
• Track personal focus pattern trends over time
• Suggest breaks when focus patterns shift toward fatigue
• Help users identify environmental factors that correlate with focus patterns

❌ Forbidden

• Diagnose ADHD or attention disorders
• Measure cognitive ability or intelligence
• Assess "focus performance" against population norms
• Make claims about productivity or work quality
• Recommend medications (stimulants, nootropics, etc.)
• Claim certainty about what the user is thinking about

Brain Signal Indicators

Primary Signals

Important Caveats

• Individual variability: Some people have naturally high/low beta
• Context matters: Beta can be elevated during anxiety, not just focus
• Multi-signal required: Never use beta alone to determine focus state

Personal Baseline Requirements

Focused State Baseline

Before the Focus Agent can provide feedback, it must learn the user's personal "focused" brain pattern:

Minimum requirements:

• 20+ sessions labeled as "focused" or "concentration tasks"
• Across different times of day
• 30-90 days to establish stable baseline

How to Build Focused Baseline

const focusedBaseline = {
  betaPower: {
    mean: 32.5,
    stdDev: 6.2,
    p25: 28.1,
    median: 32.0,
    p75: 36.8,
    observances: 45 // sessions
  },
  betaThetaRatio: {
    mean: 2.3,
    stdDev: 0.5,
    typical: 2.0-2.8
  },
  alphaPower: {
    mean: 24.1, // Typically suppressed during focus
    stdDev: 5.3
  },
  context: "User-reported focus sessions during work, coding, reading"
};

Without this baseline: Focus Agent remains silent.

Focus Pattern Detection

Detection Logic

function detectFocusPattern(eeg, user) {
  if (user.bciBaselineStatus !== 'STABLE') {
    return null; // No feedback without baseline
  }
  
  const betaRatio = eeg.betaPower / user.personalBaseline.focusedState.betaPower.median;
  const betaThetaRatio = eeg.betaPower / eeg.thetaPower;
  const alphaReduced = eeg.alphaPower < user.personalBaseline.restingState.alphaPower.median * 0.85;
  
  // Multi-signal confirmation
  const focusScore = 
    (betaRatio > 1.1 ? 30 : 0) +
    (betaThetaRatio > user.personalBaseline.focusedState.betaThetaRatio.mean - 0.5 ? 40 : 0) +
    (alphaReduced ? 30 : 0);
  
  // Require at least 2 of 3 signals for focus state
  if (focusScore >= 60) {
    return {
      state: "focused_pattern",
      confidence: focusScore / 100,
      patternMatch: "similar_to_past_focus_sessions"
    };
  }
  
  return null;
}

Temporal Requirements

Focus pattern must be sustained for 10+ seconds before triggering feedback. Brief spikes are noise.

Neurofeedback Modes

Mode 1: Active Focus Training

Context: User is explicitly practicing concentration (e.g., "focus mode" enabled)

Feedback style: Real-time, visual

Example:

{
  visual: "focus_indicator_strengthening", // Bar graph or circle intensity
  message: null, // Minimal text during active practice
  audio: "subtle_tone_increase", // Optional audio feedback
  updateFrequency: "every_2_seconds"
}

Purpose: Help user learn what focused brain patterns feel like through immediate reinforcement.

Mode 2: Pattern Observation (General Use)

Context: User is not actively practicing, but system notices focus pattern shift

Feedback style: Occasional notification

Example:

{
  title: "Focus Pattern Update",
  message: `Your brain activity shows a pattern similar to your past 
            focused sessions. Your beta waves are elevated and alpha 
            waves are reduced compared to your resting baseline.`,
  suggestion: null, // Informational only
  tone: "neutral"
}

Purpose: Increase self-awareness of natural focus state transitions.

Mode 3: Focus Fatigue Detection

Context: Brain pattern shifts away from focused state toward fatigue/distraction

Feedback style: Gentle suggestion for break

Example:

{
  title: "Pattern Shift",
  message: `Your brain activity has shifted from your typical focused 
            pattern. Your beta/theta ratio is lower than usual during 
            concentration tasks.
            
            This sometimes happens during mental fatigue.`,
  suggestion: "Consider a short break when you reach a stopping point—it might help you return to a focused state.",
  tone: "helpful_not_commanding"
}

Purpose: Suggest breaks at natural attention transition points.

Safe Language Examples

✅ Allowed Phrases

❌ Forbidden Phrases

Use Cases

Use Case 1: Concentration Practice App

Goal: Help users train sustained attention through neurofeedback

Implementation:

// User starts "focus session"
app.startFocusSession();
 
// Real-time neurofeedback
setInterval(() => {
  const pattern = detectFocusPattern(getCurrentEEG(), user);
  
  if (pattern && pattern.state === 'focused_pattern') {
    displayVisualFeedback("focus_indicator_strong");
    // User sees visual confirmation they're in focused state
  } else {
    displayVisualFeedback("focus_indicator_weak");
    // Neutral, not judgmental
  }
}, 2000); // Every 2 seconds

Safety measures:

• No numeric scores
• No comparison to others
• Visual feedback only (non-distracting)
• Personal baseline required

Use Case 2: Work Session Awareness

Goal: Help remote workers notice natural attention cycles

Implementation:

// Background monitoring during work
function checkFocusPatterns() {
  const currentPattern = classifyBrainState(getCurrentEEG(), user);
  
  if (currentPattern === 'focus_fatigue_detected') {
    // Only notify if meaningful deviation sustained for 5+ minutes
    if (sustainedFor(5 * 60)) {
      notify({
        message: "Your brain pattern has shifted from your typical focused state. A short break might help.",
        frequency_limit: "max_once_per_hour"
      });
    }
  }
}

Safety measures:

• Infrequent notifications (prevent spam)
• Suggestive, not commanding
• No judgment about productivity

Pattern Correlation (Not Causation)

The Focus Agent can observe correlations with user-tracked factors:

Allowed Correlations

// User tracks: caffeine intake, time of day, environment
const insights = {
  observation: "Your focus patterns tend to be strongest in the morning between 9-11am",
  correlation: "Sessions after coffee show 15% higher beta/theta ratio than your average",
  environmental: "When using noise-cancelling headphones, your focus patterns are 20% more stable"
};

Language: "tend to", "correlate with", "often appear when"

NOT: "caused by", "because of", "requires"

Integration with Other Contexts

Sleep and Focus

Focus patterns can be weaker after poor sleep. The agent may observe:

{
  observation: "Your focus pattern today is less stable than your recent baseline.",
  possibleContext: "Your sleep last night was shorter than usual. Focus patterns often vary with sleep quality.",
  suggestion: "If possible, lighter tasks today might feel more natural."
}

Not a diagnosis. Just pattern observation with context.

Privacy Considerations

Focus patterns can reveal:

• When someone is working vs. resting
• Attention cycles throughout the day
• Potentially sensitive work patterns

Required protections:

• No third-party access to focus pattern data
• User control over when Focus Agent is active
• Explicit consent for background monitoring
• Clear opt-out

Testing Focus Agent

Test Scenarios

1. No baseline: Should remain silent regardless of beta patterns
2. Brief beta spike: Should ignore (not sustained)
3. Sustained focus pattern: Should provide feedback if in active mode
4. Focus fatigue transition: Should suggest break if appropriate
5. Sensor artifact (extreme beta): Should recognize as sensor issue, not focus

Validation Checklist

• Requires stable personal baseline before activation
• Compares to individual baseline, not population
• Uses "pattern suggests" language, not "you are focused"
• No cognitive ability claims
• No ADHD mention or diagnosis
• Respects frequency limits (no notification fatigue)
• Multi-signal confirmation (not beta alone)

Summary

The Focus Agent:

• ✅ Recognizes personal focus patterns from baseline
• ✅ Provides neurofeedback for concentration practice
• ✅ Notices when patterns shift toward fatigue
• ✅ Helps users identify focus-supporting conditions

It does NOT:

• ❌ Diagnose ADHD or attention disorders
• ❌ Measure cognitive ability
• ❌ Claim to know what user is thinking about
• ❌ Command behavior or judge productivity

Philosophy: Pattern recognition for personal awareness, not performance evaluation.