Resting-State Functional Connectivity Correlates of Emotional Memory Control under Cognitive load in Subclinical Anxiety

This study investigates how subclinical anxiety influences the resting-state functional connectivity networks supporting the volitional control of emotional memories under cognitive load, revealing that anxiety selectively moderates prefrontal connectivity during both the suppression and recall of positive and neutral memories while distinct neural patterns underlie efficient memory control across different emotional valences.

The Gist

The Big Picture: The Brain's "Mental Tug-of-War"

Imagine your brain is a busy office. You have a Main Project (like doing your taxes) and a Side Project (like remembering a funny joke or a sad memory). Usually, you can focus on the Main Project while keeping the Side Project in the background.

But what happens if your Side Project is an intrusive memory? Maybe you're trying to study, but you can't stop thinking about an embarrassing moment from last week. Or maybe you're trying to remember a phone number, but a sad memory keeps popping up.

This study asked: How does our brain handle these intrusive memories when we are already busy? And, does having a little bit of anxiety make this "mental tug-of-war" harder?

The Experiment: The "Dual-Task" Challenge

The researchers put 47 young adults through a tricky test. Think of it like trying to juggle while riding a unicycle.

1. The Memory Task: Participants looked at pictures (some happy, some sad, some neutral). They were given a command:
   • "Think": Actively remember the picture.
   • "Don't Think": Actively try to push the picture out of your mind.
2. The Distraction: While doing this, they had to perform a second, difficult visual task (comparing the tilt of two shapes). This was the "unicycle" part—keeping their balance while juggling.

The researchers measured how well they could do the visual task while trying to remember or forget the pictures. If the memory got in the way, their performance on the visual task would drop. This drop is called "interference."

The "Anxiety" Factor

None of the participants had a diagnosed anxiety disorder. However, they all had different levels of everyday anxiety — what scientists call "subclinical anxiety." This means their anxiety symptoms were real but fell below the threshold for a clinical diagnosis. The researchers split the group into "lower anxiety" and "higher anxiety" halves and compared how their brain wiring related to their memory control performance.

The Brain Scan: Looking at the "Wiring"

After the test, everyone got an MRI scan. But they didn't do the test during the scan. Instead, they looked at the brain's Resting-State Functional Connectivity (rsFC).

The Analogy: Imagine the brain is a city with roads connecting different neighborhoods (like the Memory Neighborhood, the Control Neighborhood, and the Emotion Neighborhood).

• rsFC measures how much traffic flows between these neighborhoods when the city is "at rest" (not doing a specific task).
• If two neighborhoods have a busy, well-paved highway between them, they are "highly connected."
• If the road is closed or quiet, they are "lowly connected."

The researchers wanted to see: Does the way these roads are connected predict how well someone can control their memories?

The Key Findings

Here is what they discovered, translated into plain English:

1. The "Traffic Jam" Didn't Show Up in Behavior

Surprisingly, when looking at the test scores, there was no big difference between people who were good at "Thinking" vs. "Don't Thinking." Everyone seemed to handle the juggling act about the same way.

• The Takeaway: Just because your brain is working differently doesn't always mean you perform worse on a test. The brain might be compensating in invisible ways.

2. Different Strategies for Happy vs. Sad Memories

The brain uses different "roads" to handle different types of memories:

• For Happy Memories: The most efficient people (those who could forget happy memories best) had quieter roads between their "Control Center" (Anterior Cingulate Cortex) and their "Imagination Center" (visual areas).
  • Analogy: To stop a happy memory from popping up, the brain effectively turns down the volume on the imagination center so the memory doesn't get loud enough to distract you.
• For Sad Memories: The most efficient people had louder, busier roads between their "Attention Center" (Supramarginal Gyrus) and their "Visual Processing Center."
  • Analogy: To stop a sad memory, the brain doesn't just turn down the volume; it actively redirects traffic to look at something else, like a security guard shouting, "Look over here!" to distract you from the sad thought.

3. The "Anxiety" Effect: Over-Engineering the System

This was the most interesting part. Anxiety acted like a moderator or a filter.

• The High-Anxiety Group: When these people tried to forget happy memories, their brains showed increased traffic between the "Control Center" and the "Medial Frontal Cortex."
  • Analogy: Imagine a driver who is a bit anxious. When they try to ignore a distraction, they don't just gently steer away; they grip the steering wheel so hard that the whole car shakes. They are using extra effort and extra brain power to do the same job that a calm person does easily. They are "over-driving" the system.
• The Low-Anxiety Group: They showed the opposite pattern. They didn't need to grip the wheel so hard.

4. The "Memory vs. Forgetting" Switch

When the researchers compared "Remembering" vs. "Forgetting," they found that Forgetting required a stronger connection between the Hippocampus (the brain's library) and the Thalamus (the brain's relay station).

• Analogy: To successfully "delete" a file from your computer, you need a very strong, direct cable between the hard drive and the trash can. To "open" a file, you just need a standard connection. Forgetting is an active, high-energy process.

Why Does This Matter?

This study tells us that anxiety changes the blueprint of our brain's network, even if we aren't clinically diagnosed with a disorder.

• For the anxious brain: Controlling thoughts is like driving a car with a heavy foot on the gas. It works, but it burns more fuel (mental energy).
• The Good News: Because we can see these specific "roads" (connectivity patterns), scientists can eventually develop better therapies. Instead of just telling anxious people to "calm down," we might be able to train their brains to take a more efficient route, using less energy to achieve the same result.

In a Nutshell

Your brain is a network of highways. When you try to ignore a memory, your brain builds a detour. If you have a bit of anxiety, your brain builds a massive, over-engineered detour that uses a lot of energy. This study mapped those detours, showing us exactly how anxiety changes the way we think and forget.

Technical Summary

1. Problem Statement and Research Gap

Volitional memory control—the ability to intentionally suppress interfering memories or recall goal-relevant ones—is fundamental to adaptive cognition. While previous research has established the neural mechanisms of Suppression (inhibitory control) and Recall (retrieval) largely in isolation, two critical gaps remain:

1. Concurrent Cognitive Load: It is unclear how these mechanisms operate when individuals must simultaneously manage a secondary, demanding cognitive task (dual-task interference).
2. Subclinical Anxiety: Most studies focus on clinical anxiety disorders. The neural and behavioral correlates of emotional memory control in individuals with subclinical anxiety (elevated symptoms below diagnostic thresholds) under cognitive load are poorly characterized, despite this group representing a significant at-risk population.

The study aims to investigate how subclinical anxiety modulates the relationship between dual-task memory-linked interference (cognitive control efficiency) and Resting-State Functional Connectivity (rsFC) during the directed forgetting of emotionally valenced memories.

2. Methodology

Participants

• Sample: 47 healthy young adults (14 females, 33 males; mean age ~21.7 years).
• Inclusion Criteria: No history of neurological/psychiatric disorders, normal vision, and no prescription drugs affecting the nervous system.
• Anxiety Assessment: Participants were screened using the PHQ-9 (depression) and the Fear Affect scale (NIH Toolbox) for anxiety. Participants were split into "low" and "high" anxiety subsets based on the median score of the sample (Median = 14).

Experimental Design

The study employed a two-day protocol:

1. Day 1 (Behavioral Task): An Item-Method Directed Forgetting paradigm combined with a concurrent Visual Working Memory (WM) task.
   • Stimuli: Emotional images (Negative, Neutral, Positive) from the NAPS database.
   • Instructions: After viewing an image, a cue indicated either Recall (green) or Suppress (red) the memory.
   • Dual-Task: Immediately following the cue, participants performed a visual WM task (judging the orientation of Gabor patches).
   • Metric: Balanced Integration Score (BIS) was calculated as $Z_{accuracy} - Z_{RT}$. Higher BIS indicates better cognitive control (less interference from memory processes on the concurrent WM task).
   • Validation: A recognition task confirmed participants followed instructions (higher sensitivity for Recall than Suppression).
2. Day 2 (fMRI): A 11-minute Resting-State fMRI (rs-fMRI) scan (eyes open) was acquired to measure intrinsic functional connectivity.

Neuroimaging Analysis

• Preprocessing: Performed using the CONN toolbox (realignment, slice-timing, denoising, normalization to MNI space, smoothing).
• Seed-to-Voxel Analysis: A priori Regions of Interest (ROIs) were selected based on systematic literature reviews and NeuroQuery meta-analyses. Key seeds included the Anterior Cingulate Cortex (ACC), Hippocampus, Supramarginal Gyrus, Inferior Frontal Gyrus, and Superior Frontal Gyrus.
• Statistical Models:
  • Equation 2: Association between BIS and rsFC.
  • Equation 3: Moderation analysis (BIS $\times$ Anxiety) to test if anxiety severity alters brain-behavior relationships.
  • Equation 4: Direct comparison of Suppression vs. Recall connectivity.
• Correction: Voxel-wise threshold $p < 0.001$ (uncorrected) and cluster-level $p < 0.05$ (FWE-corrected).

3. Key Results

Behavioral Findings

• No Dual-Task Differences: There were no significant differences in the Balanced Integration Score (BIS) between Recall and Suppression conditions, nor across emotional valences (Negative, Neutral, Positive).
• Anxiety & Behavior: No significant correlation was found between anxiety levels and behavioral performance (BIS) across conditions.
• Interpretation: The concurrent WM load did not differentially impact the efficiency of Recall vs. Suppression in this sample, suggesting comparable cognitive resource allocation or high individual variability in WM capacity.

Resting-State Functional Connectivity (rsFC) Findings

A. Suppression-Linked Cognitive Control (Main Effects)

• Positive Memories: Efficient suppression correlated with reduced rsFC between:
  • ACC (Salience Network) and posterior midline/visual regions (Intracalcarine Cortex, Precuneus, Supracalcarine Cortex).
  • Right Hippocampus and the Frontal Pole.
  • Implication: Efficient suppression involves downregulating connectivity between control regions and areas involved in perceptual reinstatement and internal mental states.
• Negative Memories: Efficient suppression correlated with increased rsFC between:
  • Right Posterior Supramarginal Gyrus (pSMG) and the Left Angular Gyrus and Superior Lateral Occipital Cortex.
  • Implication: Suppression of negative content relies on attentional reorienting and perceptual control networks.

B. Anxiety Moderation Effects

• Suppression of Positive Memories: Anxiety significantly moderated the relationship between BIS and rsFC.
  • High Anxiety: Showed a positive correlation between BIS and connectivity between the Right Superior Frontal Gyrus (SFG) and Medial Frontal Cortex/Anterior Cingulate.
  • Low Anxiety: Showed a negative correlation.
  • Implication: Individuals with higher anxiety engage compensatory prefrontal control mechanisms more intensely as suppression efficiency increases.
• Recall of Positive/Neutral Memories: Anxiety moderated connectivity involving the Left Inferior Frontal Gyrus (pars triangularis).
  • High Anxiety: Reduced connectivity between IFG and distributed regions (temporal, parietal, frontal).
  • Low Anxiety: Positive connectivity.
  • Implication: High anxiety is associated with reduced engagement of networks typically involved in elaborating and retrieving positive/neutral memories.

C. Differential Connectivity (Suppression vs. Recall)

• Direct comparison revealed that Suppression of positive memories was associated with increased rsFC between the Right Hippocampus and the Thalamus/Caudate compared to Recall.
• This suggests that suppressing positive memories under load requires heightened hippocampal-thalamic communication, possibly to sustain regulation of reactivation.

4. Key Contributions

1. Network-Level Specificity: The study delineates distinct rsFC profiles for Suppression vs. Recall, showing that these processes rely on dissociable large-scale networks (e.g., Salience-Parietal for positive suppression vs. pSMG-Parietal for negative suppression).
2. Valence-Dependent Mechanisms: It demonstrates that the neural correlates of memory control are not uniform but vary significantly based on emotional valence (Positive vs. Negative).
3. Anxiety as a Moderator: It provides evidence that subclinical anxiety does not necessarily degrade behavioral performance but selectively reshapes the neural architecture supporting memory control. Specifically, anxiety drives compensatory prefrontal engagement during suppression and alters connectivity during recall.
4. Dual-Task Context: It extends the understanding of memory control to ecologically valid scenarios involving concurrent cognitive load, revealing that intrinsic connectivity patterns predict control efficiency even when behavioral differences are absent.

5. Significance and Limitations

• Significance: The findings support a dimensional view of anxiety, suggesting that subclinical variations in symptom severity are sufficient to alter the intrinsic functional organization of brain networks involved in emotional regulation. This offers potential biomarkers (specific rsFC patterns) for early intervention or mechanistic studies.
• Limitations:
  • Sample Demographics: The sample was heavily male-skewed (~3:1), limiting generalizability given known sex differences in anxiety and emotional processing.
  • Cross-Sectional Design: Causality cannot be inferred; longitudinal studies are needed to determine if these connectivity patterns precede or result from anxiety symptoms.
  • Behavioral Null Effects: The lack of behavioral differences suggests the dual-task load may not have been sufficiently demanding to elicit performance deficits, though it successfully revealed neural differences.

Conclusion

The study successfully maps the intrinsic functional connectivity correlates of volitional emotional memory control under cognitive load. It establishes that while behavioral performance may appear stable across anxiety levels, the underlying neural networks are significantly modulated by subclinical anxiety, particularly involving prefrontal-limbic and fronto-parietal circuits. These findings provide a refined neurobiological framework for understanding how anxiety influences the regulation of intrusive thoughts in everyday, cognitively demanding contexts.