Sleep neural code perpetuates the evolving negativity bias under stress

This study demonstrates that sleep causally perpetuates stress-induced negativity bias by orchestrating the co-reactivation of stress-tagged cellular ensembles in the hippocampus, which generates persistent negative schema-like representations that distort future memory recall and emotional processing.

The Gist

The Big Picture: Why Does a Bad Day Ruin a Week?

Have you ever had a really terrible day, and even though the next day was fine, you still felt like everything was going wrong? You might look at a neutral situation and think, "Oh no, this is going to be a disaster," even if there's no evidence for it. Psychologists call this negativity bias. It's like your brain has a broken filter that only lets the bad stuff through, making you overgeneralize your fears.

This paper asks a simple but profound question: How does the brain get stuck in this negative loop, and why does it keep happening even after the stress is gone?

The answer, according to this study, is sleep. Specifically, the way your brain replays and reorganizes memories while you are unconscious is what cements that negative outlook.

---

The Experiment: The "Bully" and the "Safe Zone"

To figure this out, the scientists used mice (because their brains work similarly to ours in many ways).

1. The Bad Experience (The Bully): They subjected some mice to "social defeat stress." Imagine a mouse being bullied by a bigger, aggressive mouse every day for six days. This creates a stable, negative emotional state. The mice became anxious and avoided other mice.
2. The Test (The New Situation): Later, they put these bullied mice in a new situation: a fear test where they got a tiny, harmless electric shock.
3. The Result: The bullied mice didn't just remember the shock; they became super scared of it. Even more importantly, they got scared of new, safe places that looked slightly different. Their brains had decided, "Everything is dangerous now."

The Discovery: The Brain's "Negative Schema"

The researchers looked inside the mice's brains (specifically the hippocampus, the memory center) using a special camera that lets them see neurons lighting up like tiny stars.

They found that the stressed mice developed something called a "Negative Schema."

• The Analogy: Imagine your brain is a library. Usually, books are organized by topic: "Cooking," "History," "Sports."
• The Problem: When the mice were stressed, their brain started creating a new, sticky category called "DANGER."
• The Glitch: Even when a new book (a new experience) came in that was actually about "Cooking" (a safe situation), the brain's librarian (the neurons) automatically stamped it with the "DANGER" label because it felt vaguely similar to the bad stuff. The brain wasn't just remembering the specific bullying; it was creating a general rule that everything is bad.

The Plot Twist: Sleep is the Editor

Here is the most surprising part. The researchers watched what happened when the mice were awake versus when they were asleep.

• While Awake: The brain was busy reacting to the immediate moment.
• While Asleep: The brain started doing something magical and terrifying. It began replaying the memories of the bullying and the new fear test together.

The Analogy: Think of your brain as a movie editor.

• During the day (Awake): The editor is just filming the scenes as they happen.
• At night (Sleep): The editor goes into the editing room. They take the footage of the "Bully" and the footage of the "New Safe Room" and splice them together. They mix the audio tracks so that the sound of the bully is playing over the footage of the safe room.

By the time the mouse wakes up, the brain has convinced itself that the "Safe Room" and the "Bully" are the same thing. The negative feeling has been generalized.

The "Off Switch" Experiment

To prove that sleep was the cause of this negativity, the scientists did something very cool. They used a "remote control" (optogenetics) to silence specific brain cells in the stressed mice only while they were sleeping.

• The Result: When they silenced the brain cells during sleep, the mice stopped being overly negative the next day. They could tell the difference between a scary situation and a safe one.
• The Control: When they silenced the same cells while the mice were awake, nothing changed. The negativity bias remained.

The Takeaway: It's not just about having the bad memory; it's about how your brain processes that memory while you sleep. If you interrupt that processing, you can stop the negativity from spreading to new situations.

Why This Matters for Us

This study gives us a biological explanation for why we sometimes feel stuck in a negative mindset after a trauma or a stressful period.

1. Sleep is Active, Not Passive: Sleep isn't just "shutting down." It's a time when your brain is actively rewriting your emotional reality.
2. The Danger of Rumination: If you are stressed, your brain might be over-editing your memories at night, turning small worries into giant fears.
3. A New Hope for Therapy: This suggests that treatments for anxiety and depression might need to focus on sleep quality and sleep architecture. If we can help the brain process negative memories correctly during sleep (perhaps by preventing that "over-generalization" splicing), we might be able to break the cycle of negativity bias.

In short: Your brain is like a DJ at night. If you've had a bad day, it might start mixing all your songs into a sad, scary remix. But if we can learn to control the DJ's playlist during sleep, we might be able to wake up feeling like the world is a little less scary.

Technical Summary

1. Problem Statement

Negativity bias is a core cognitive distortion in mood disorders (e.g., depression, anxiety) where individuals interpret neutral or ambiguous experiences through a negative lens, leading to rumination and overgeneralization. While psychological models (e.g., Beck's cognitive theory) propose that negative "schemas" drive this bias, and neuroimaging links it to default mode network activity, the biological mechanisms underlying how these semantic-like negative representations are generated, sustained, and propagated across time remain unknown. Specifically, it is unclear how offline brain states (sleep) contribute to the consolidation and evolution of these negative emotional schemas.

2. Methodology

The authors employed a multi-modal approach combining behavioral paradigms, longitudinal in vivo calcium imaging, chemogenetics, and closed-loop optogenetics in mice.

• Behavioral Models:

  • Repeated Social Defeat Stress (SoD): Induced a stable negative emotional state in male mice over 6 days.
  • Contextual Fear Conditioning (CFC) & Trace Fear Conditioning (TFC): Used to test memory recall and the generalization of fear to novel contexts/stimuli.
  • Social Interaction (SI) & Elevated Plus Maze (EPM): Quantified social avoidance and anxiety levels.
  • Positive Control: Female Social Interaction (FSI) was used to induce a positive state for comparison.

• Neural Recording & Imaging:

  • Longitudinal Ca²⁺ Imaging: Used AAV9-jGCaMP7b injected into the dorsal hippocampal CA1 region of cFos-tTA transgenic mice. This allowed tracking of the same neuronal ensembles across multiple days and distinct emotional experiences (SoD, CFC, SI).
  • EEG/EMG: Recorded simultaneously to score sleep stages (Wake, NREM, REM) in real-time.

• Perturbation Techniques:

  • Chemogenetics (DREADDs): Used cFos-tTA mice to tag ensembles active during SoD with inhibitory hM4Di receptors. Administration of the agonist J60 silenced these specific ensembles.
  • Closed-Loop Optogenetics: Tagged SoD-ensembles with the inhibitory opsin ArchT3.0. A wireless system (Teleopto) coupled with real-time EEG/EMG scoring allowed for the selective silencing of these ensembles only during specific sleep states (NREM, REM) or wakefulness immediately following a new learning event (CFC).

• Computational Analysis:

  • PCA/ICA: Used to extract independent neuronal ensembles (cell assemblies) from Ca²⁺ data.
  • Co-reactivation Analysis: Quantified the simultaneous reactivation of distinct ensembles (e.g., SoD-related vs. CFC-related) during offline periods.
  • Population Vector (PV) & MDS: Analyzed the geometric similarity of neural representations across sessions.
  • SVM Decoding: Trained Support Vector Machines to determine if specific sleep-active neuronal subsets could decode future aversive stimuli or freezing behaviors.

3. Key Contributions

• Discovery of Semantic Negative Schemas: Demonstrated that the dorsal hippocampus generates "schema-like" neuronal representations that encode the semantic concept of negativity (generalized negative affect) rather than just specific episodic details.
• Causal Role of Sleep: Established that sleep, specifically the interaction between NREM and REM, is causally required to perpetuate negativity bias. Silencing stress-tagged ensembles during sleep, but not during wakefulness, prevents future negative bias.
• Mechanism of Overgeneralization: Revealed that sleep orchestrates the co-reactivation of distinct emotional experiences (past stress and new fear), integrating them into a unified negative representation that biases future perception of novel stimuli.
• Predictive Coding: Showed that sleep-active neuronal populations can decode future aversive stimuli and freezing behaviors, suggesting sleep pre-defines the emotional interpretation of future conscious experiences.

4. Key Results

A. Induction of Negativity Bias

• Mice subjected to 6 days of SoD exhibited social avoidance and increased anxiety (EPM).
• Crucially, these mice showed enhanced freezing during CFC recall and, significantly, elevated freezing in novel contexts and to novel tones (generalization), indicating a bias where new experiences are interpreted negatively.
• A positive emotional state (FSI) produced the opposite effect (reduced freezing), confirming the state-dependency of the bias.

B. Emergence of Schema-Like Representations

• Longitudinal imaging revealed that in SoD mice, specific ensembles in the dorsal hippocampus became active across different negative experiences (SoD days, SI with aggressive mice, CFC recall).
• These "CFC recall-specific ensembles" showed increased activity during social avoidance (opposite corner stays) and freezing, correlating with the degree of future bias.
• Functional connectivity analysis showed increased coordination between social avoidance-related and fear-related cell subtypes specifically during the negative state.

C. Sleep Dynamics and Co-reactivation

• Offline Processing: During REM sleep, the activity of schema-like ensembles increased significantly in SoD mice after new negative experiences, unlike in neutral mice.
• Co-reactivation: During NREM sleep, distinct ensembles representing past stress (SoD) and new fear (CFC) showed robust co-reactivation in SoD mice, but not in neutral mice or during wakefulness.
• This NREM co-reactivation was followed by REM activation of the schema, suggesting a two-step process: NREM extracts the gist (integration), and REM stabilizes the schema.

D. Causal Manipulation (Optogenetics & Chemogenetics)

• Chemogenetic Silencing: Silencing SoD-tagged ensembles reduced baseline anxiety and prevented the overgeneralization of fear to novel contexts.
• Closed-Loop Optogenetics: Silencing SoD-tagged ensembles during sleep (NREM, REM, or both) immediately after CFC training completely abolished the enhanced freezing in future recall and novel contexts.
• Critical Finding: Silencing the same ensembles during wakefulness had no effect on future bias. This proves that sleep-specific dynamics, not just the presence of the ensembles, are causal for the perpetuation of negativity.

E. Decoding Future Bias

• SVM decoders trained on sleep-active SoD cells (those reactivated during sleep) successfully predicted CS+ trials and freezing behavior in future recall sessions, including responses to novel ambiguous tones.
• Sleep-non-active cells failed to decode these future behaviors.
• Population vector analysis showed that neural representations during post-SoD sleep were geometrically closer to future aversive recall states than to pre-SoD states, confirming that sleep reshapes the neural code toward negativity.

5. Significance

• Biological Basis for Cognitive Theory: This study provides the first biological evidence for the existence of "negative schemas" in the brain, showing how episodic memories are abstracted into semantic-like representations of negativity in the dorsal hippocampus.
• Sleep as a Cognitive Platform: It redefines the role of sleep in emotional health. Rather than just consolidating memories, sleep actively constructs and updates emotional interpretations. In stress, this mechanism malfunctions, leading to a self-perpetuating cycle of negative overgeneralization.
• Therapeutic Implications: The findings suggest that targeting sleep-specific neural dynamics (e.g., disrupting the co-reactivation of negative ensembles during NREM/REM) could be a novel therapeutic strategy for mood disorders, potentially offering a mechanism to break the cycle of rumination and negative bias that pharmacological treatments and CBT aim to address.
• Mechanism of Generalization: It elucidates how the brain transitions from specific fear memories to generalized anxiety, identifying the specific sleep-dependent neural computations responsible for this pathological broadening of emotional boundaries.