2nd/10_Wiki/Topics/AI_and_ML/Default Mode Network (DMN).md

id, title, category, status, canonical_id, aliases, duplicate_of, source_trust_level, confidence_score, verification_status, tags, raw_sources, last_reinforced, github_commit, tech_stack

id	title	category	status	canonical_id	aliases	duplicate_of	source_trust_level	confidence_score	verification_status	tags	raw_sources	last_reinforced	github_commit	tech_stack
wiki-2026-0508-default-mode-network	Default Mode Network (DMN)	10_Wiki/Topics	verified	self	DMN default mode network mind wandering rumination task-negative sleep consolidation	none	A	0.88	applied	neuroscience dmn mind-wandering creativity depression sleep-consolidation brain-network		2026-05-10	pending	language applicable_to neuroscience Productivity Creativity Mental Health AI Memory Design

Default Mode Network (DMN)

매 한 줄

"매 idle 의 brain 의 가장 busy". 매 mPFC + PCC + angular gyrus 의 task-negative network. 매 self-reflection + 매 social cognition + 매 creativity 의 source. 매 over-active = 매 rumination (depression). 매 modern AI: 매 background memory consolidation 의 inspiration.

매 핵심 region

• Medial Prefrontal Cortex (mPFC): 매 self-reference.
• Posterior Cingulate Cortex (PCC): 매 hub.
• Angular Gyrus: 매 association.
• Hippocampus: 매 memory replay.
• Temporal pole: 매 social knowledge.

매 function

1. Self-referential thought: "Who am I?"
2. Social cognition / ToM: 매 others 의 mental state 의 simulate.
3. Autobiographical memory: 매 past recall.
4. Future imagination / planning.
5. Creative incubation: 매 random connection.
6. Mind wandering: 매 default state.

Task-positive vs Task-negative

• TPN (Task-Positive): 매 focused work — DMN 의 suppress.
• DMN (Task-Negative): 매 rest / wandering — TPN 의 suppress.
• 매 anti-correlated.

매 functional balance

• Healthy: 매 flexible switching.
• Rumination (depression): 매 DMN over-active.
• ADHD: 매 DMN suppression 의 fail.
• Autism: 매 DMN 의 underconnected.
• Schizophrenia: 매 DMN 의 abnormal.

매 modulation

• Mindfulness / meditation: 매 DMN 의 quiet.
• Psychedelic (psilocybin): 매 ego dissolution = DMN 의 disrupt.
• Sleep: 매 DMN-like activity (consolidation).
• Exercise: 매 BDNF + 매 DMN 의 modulate.
• Boredom / nature: 매 DMN 의 activate.

매 productivity 의 응용

• Creative break: 매 walking, shower.
• Active rest: 매 mindfulness > 매 social media (매 same wandering 의 X).
• Pomodoro + 매 break.
• Sleep: 매 problem 의 forget → 매 wake 의 insight.

매 modern AI 의 inspiration

• Sleep replay (RL): 매 hippocampal replay 의 memory.
• Latent dreaming: 매 model 의 internal simulation.
• DPO / RLHF: 매 quiet phase + 매 reflection.
• Background processing: 매 LLM 의 idle 의 memory consolidate.

💻 패턴 (응용)

Cognitive worker 의 DMN-friendly schedule

def dmn_friendly_day():
    return [
        ('06:30-07:00', 'Wake + sunlight (no phone)', 'DMN active'),
        ('07:00-07:30', 'Walk in nature', 'DMN + creative incubation'),
        ('07:30-08:00', 'Breakfast + reflective journal', 'DMN'),
        ('08:30-10:00', 'Deep work (TPN)', 'DMN suppressed'),
        ('10:00-10:30', 'Walk break (no phone, no podcast)', 'DMN — incubate'),
        ('10:30-12:00', 'Deep work', 'TPN'),
        ('12:00-13:00', 'Lunch + slow walk', 'DMN'),
        ('13:00-14:30', 'Deep work', 'TPN'),
        ('14:30-15:00', 'Power nap or meditation', 'DMN consolidate'),
        ('15:00-17:00', 'Light tasks', ''),
        ('22:30', 'Wind down (no screen 60 min)', 'DMN prep'),
    ]

Mindfulness session (DMN modulator)

def mindfulness_3min():
    return [
        ('1 min', 'Awareness — what is here?', 'DMN observe'),
        ('1 min', 'Breath anchor', 'DMN suppress'),
        ('1 min', 'Open awareness', 'DMN with awareness'),
    ]

Sleep consolidation (replay-inspired RL)

def replay_buffer_consolidation(buffer, model):
    """매 sleep / idle phase 의 replay."""
    # 매 매 hour 의 1× of recent 의 rehearse
    samples = buffer.sample(n=batch_size, prefer_recent=True)
    
    # 매 with light noise (creative recombine)
    for s in samples:
        s_perturbed = add_small_noise(s)
        model.gradient_step(s_perturbed)

Rumination detector (mental health)

def detect_rumination(journal_entries, days=7):
    """매 negative repetitive thought 의 detect."""
    recent = journal_entries[-days:]
    sentiment_neg_count = sum(1 for e in recent if sentiment(e) < -0.3)
    
    # 매 same theme 의 repetition
    themes = [extract_theme(e) for e in recent]
    repeated = Counter(themes).most_common(1)[0][1]
    
    if sentiment_neg_count > days * 0.6 and repeated > days * 0.5:
        return 'WARN: rumination risk — consider professional support'
    return None

Creative incubation (deliberate wandering)

def creative_incubation(problem, duration_min=20):
    """매 problem 의 conscious 의 release + DMN 의 work."""
    return [
        f'1. Define problem clearly (5 min)',
        f'2. Set aside ({duration_min - 5} min):',
        f'   - Walk in nature',
        f'   - Take shower',
        f'   - Light sleep / meditation',
        f'   - Repetitive low-stim activity (folding laundry)',
        f'3. Return + capture insights without forcing',
    ]

LLM "DMN" mode (background memory)

class LLMWithMemory:
    def __init__(self, model):
        self.model = model
        self.episodic_buffer = []
    
    async def task_mode(self, query):
        return await self.model.generate(query)
    
    async def dmn_mode(self):
        """매 idle 의 background consolidation."""
        if not self.episodic_buffer: return
        
        # 매 random recent episode 의 reflect
        episode = random.choice(self.episodic_buffer[-100:])
        reflection = await self.model.generate(
            f"Reflect on this past interaction:\n{episode}\n\n"
            f"Extract 1-3 lessons or patterns. Brief."
        )
        save_long_term(reflection)

Anti-DMN 의 toxic habit

DMN_KILLER = [
    'social media scroll (passive + comparing)',
    'doom-scrolling news',
    'multi-tasking video',
    'endless email check',
]

DMN_FRIENDLY = [
    'walk in nature',
    'shower',
    'mindless physical task (laundry, cooking simple)',
    'meditation',
    'journal',
    'reading fiction (no notes)',
]

매 결정 기준

상황	Activity
Stuck on problem	DMN-friendly walk / nap
Creative work needed	Schedule incubation time
Mental fatigue	Mindfulness or nature
Rumination risk	CBT + meditation + exercise
Productivity	TPN block + DMN break
Insomnia	Anti-rumination CBT-I

기본값: 매 deep work + 매 DMN-friendly break + 매 sleep + 매 mindfulness.

🔗 Graph

• 부모: Neuroscience · Cognitive-Science
• 변형: Mind-Wandering · Rumination · Sleep-Consolidation · Theory-of-Mind
• 응용: Mindfulness · Creative-Incubation · Pomodoro
• Adjacent: Brain-Derived Neurotrophic Factor (BDNF) · Bayesian-Brain-Hypothesis · Cognitive-Therapy-in-CBT · Cognitive Reserve Theory · Bioenergetics

🤖 LLM 활용

언제: 매 productivity routine. 매 creative work design. 매 mental health prevention. 매 LLM background memory design. 언제 X: 매 specific clinical (의사 / therapist).

❌ 안티패턴

• No DMN time (constant TPN): 매 burnout + 매 creative block.
• Social media as "rest": 매 DMN의 X — 매 same task-engagement.
• Rumination 의 ignore: 매 depression risk.
• Boredom 의 fear: 매 DMN 의 valuable.
• Sleep 의 cut: 매 consolidation X.

🧪 검증 / 중복

• Verified (Raichle DMN 2001, Buckner DMN reviews, Carhart-Harris psychedelic + DMN).
• 신뢰도 A.
• Related: Brain-Derived Neurotrophic Factor (BDNF) · Bayesian-Brain-Hypothesis · Cognitive-Therapy-in-CBT · Cognitive Reserve Theory · Bioenergetics.

🕓 Changelog

날짜	변경
2026-05-08	Phase 1
2026-05-10	Manual cleanup — region + function + 매 schedule / mindfulness / replay / LLM DMN code