Brain volumetric changes in the general population following the COVID-19 outbreak and lockdown: The intense experience induced transient volumetric changes in brain regions commonly associated with stress & anxiety

Brain volumetric changes in the general population following the COVID-19 outbreak and lockdown. Tom Salomon et al. NeuroImage, June 26 2021, 118311. https://doi.org/10.1016/j.neuroimage.2021.118311

Abstract: The coronavirus disease 2019 (COVID-19) outbreak introduced unprecedented health-risks, as well as pressure on the economy, society, and psychological well-being due to the response to the outbreak. In a preregistered study, we hypothesized that the intense experience of the outbreak potentially induced stress-related brain modifications in the healthy population, not infected with the virus. We examined volumetric changes in 50 participants who underwent MRI scans before and after the COVID-19 outbreak and lockdown in Israel. Their scans were compared with those of 50 control participants who were scanned twice prior to the pandemic. Following COVID-19 outbreak and lockdown, the test group participants uniquely showed volumetric increases in bilateral amygdalae, putamen, and the anterior temporal cortices. Changes in the amygdalae diminished as time elapsed from lockdown relief, suggesting that the intense experience associated with the pandemic induced transient volumetric changes in brain regions commonly associated with stress and anxiety. The current work utilizes a rare opportunity for real-life natural experiment, showing evidence for brain plasticity following the COVID-19 global pandemic. These findings have broad implications, relevant both for the scientific community as well as the general public.

3. Discussion

Our study demonstrates that volumetric change patterns in the brain occurred following the COVID-19 initial outbreak period and restrictions in a sample of healthy participants, who were not somatically affected by the pandemic. While previous studies demonstrated brain plasticity using T1-weighted MRI following planned interventions (Maguire et al., 2000, Jung et al., 2013, Draganski et al., 2004), the current work outstands in its unique demonstration of stark structural brain plasticity following a major real-life event.

Our findings show neural changes that were not caused directly due to COVID-19 infection, but rather related to the societal effect, further resonating the mental contagiousness aspect of the COVID-19 pandemic (Valenzano et al., 2020). We show volumetric increase in gray matter in the amygdalae, putamen, and ventral anterior temporal cortices. The changes in the amygdalae showed a temporal-dependent effect, related to the time elapsed from lockdown but not the duration from the baseline scan. It should be noted that although lockdown restrictions had initially reduced infection rates in Israel, just one month after the lockdown was lifted, the number of infected cases started to rise again and reached higher number of active infected cases by the end of data collection, compared with the peak numbers during the actual lockdown period (approximately 2,000 daily new cases by the end of July versus under 750 new daily cases during the peak of the lockdown period in April (Israel Ministry of Health 2020), see detailed timeline in the methods section and Figure 4). This suggests that the effects observed in the current study are less likely to be attributed to the concrete health risks of contracting the virus, but rather to the first wave of the outbreak, characterized with perceived uncertainty and substantial unexpected changes in everyday life.

Figure 4. Study timeline and outbreak data

On February 21st, 2020, the first COVID-19 case in Israel was recorded. Daily new cases were smoothed using 7-days moving average. Data were retrieved and modified based on the Israeli Ministry of Health reports (Israel Ministry of Health 2020, Max et al., 2020). A lockdown was issued on March 25th, which was gradually released until the removal of the 100-meter restriction on May 1st, marking lockdown onset and relief, respectively (shorter vertical dashed line). MRI data of the test group were collected between May 10th to July 29th (longer vertical dashed line). Red bars on top represent the number of participants scanned for the study each day.

Examining the contribution of study features such as volumetric measurements at baseline, the initial study, and scan angle, revealed that the volumetric change effects in the bilateral amygdalae and temporal cortical ROIs, were mostly stable. Although some confirmatory analysis with confounding covariates slightly reduced the significance of the group-time interaction effect, this decrease was relatively small (with significant results before FDR correction), and more importantly, the confounding factor were not significant in any of the models. Thus, it is unlikely that a confound related to the study design could account for the volumetric change effect. Changing the analysis pipeline from surface-based to voxel-based morphometry, resulted in non-significant effect in the Putamen; thus, suggesting that the effect in these nuclei might be susceptible to differences in analysis pipeline. Putatively, the results in these regions change due to different segmentation of the nuclei, registration or smoothing. Therefore, conclusions regarding volumetric change in the Putamen should be more reserved.

The current literature regarding volumetric changes in the amygdala following stressful events, and especially real-life events, is quite limited. Some studies found evidence in agreement with our results, such as one work showing decreased amygdala volume was associated with greater stress reduction following mindfulness training (Hölzel et al., 2009); while others found evidence in the opposite direction, such as one study which found smaller amygdala volumes within participants who were in closer proximity to the World Trade Center during 9/11 events (Ganzel et al., 2008), and overall meta-analyses approach often showing contradicting evidence regarding amygdala volumetric difference within population associated with stress such as post-traumatic stress disorder (PTSD) and generalized stress disorder (O'Doherty et al., 2015, Duval et al., 2015). Our results, showing a gradual decline of the volumetric change effect as a function of TFL, could provide a potential insight into these inconclusive patterns. It is possible that without time-locking to a strong external event, volumetric change effect would be more difficult to detect. This point highlights the uniqueness of our study that included a repeated session design within with a real-world event.

The current study was in many aspects unplanned; therefore, we are left with only partial answers as to which specific behavioral or cognitive impacts of the COVID-19 outbreak led to the neural changes observed in the healthy participants that took part in our study. The involvement of the amygdala may suggest that stress and anxiety could be the source of the observed phenomenon, due to its well-recorded functional and structural associations (Ganzel et al., 2008, Hölzel et al., 2009, Rogers et al., 2009, Schienle et al., 2011, Mochcovitch et al., 2014, Bryant et al., 2008, Stevens et al., 2017). Nevertheless, it is hard to draw clear conclusions as many aspects of life have changed in this time period, and could have potentially affected different regions in the brain – from limiting social interactions, increased financial stress, changes in physical activity, work routine, and many more. The limited behavioral data collected in the current study did not provide a strong connection to the imaging results, and thus future work could try to better address the complex brain-behavioral associations in this real-life experience.

Furthermore, as our study only examined T1-weighted anatomical scans, we are limited in our scope to gross-anatomy macroscale changes. Imaging research using additional imaging methods such as diffusion tensor imaging (DTI) and functional MRI (fMRI), showed that neural plasticity processes are often characterized by changes of microstructural scale, commonly expressed in the white matter (Sagi et al., 2012, Scholz et al., 2009, Sampaio-Baptista et al., 2013, Steele and Zatorre, 2018) and functional neural activity (Brodt et al., 2018), which were not examined here. Further research combining both more extensive behavioral and imaging measurements might be able to link brain modification with specific behavioral manifestations of COVID-19 outbreak.

Despite these limitations, our findings show that healthy young adults, with no records of mental health issues, were deeply affected by the outbreak of COVID-19. These findings are both ground-breaking in showing brain plasticity of subcortical regions following real-life external event, as well as in revealing an additional impact of the COVID-19 on the well-being of the general public. Our results emphasize the impact of widescale societal changes and suggest that when forming such changes, one should take into consideration the indirect impact on the general well-being of the population, alongside the efficacy of the societal changes.