Increased sensitivity to strong perturbations in a whole-brain model of LSD

Abstract

Lysergic acid diethylamide (LSD) is a potent psychedelic drug, which has seen a revival in clinical and pharmacological research within recent years. Human neuroimaging studies have shown fundamental changes in brain-wide functional connectivity and an expansion of dynamical brain states, thus raising the question about a mechanistic explanation of the dynamics underlying these alterations. Here, we applied a novel perturbational approach based on a whole-brain computational model, which opens up the possibility to externally perturb different brain regions in silico and investigate differences in dynamical stability of different brain states, i.e. the dynamical response of a certain brain region to an external perturbation. After adjusting the whole-brain model parameters to reflect the dynamics of functional magnetic resonance imaging (fMRI) BOLD signals recorded under the influence of LSD or placebo, perturbations of different brain areas were simulated by either promoting or disrupting synchronization in the regarding brain region. After perturbation offset, we quantified the recovery characteristics of the brain area to its basal dynamical state with the Perturbational Integration Latency Index (PILI) and used this measure to distinguish between the two brain states. We found significant changes in dynamical complexity with consistently higher PILI values after LSD intake on a global level, which indicates a shift of the brain's global working point further away from a stable equilibrium as compared to normal conditions. On a local level, we found that the largest differences were measured within the limbic network, the visual network and the default mode network. Additionally, we found a higher variability of PILI values across different brain regions after LSD intake, indicating higher response diversity under LSD after an external perturbation. Our results provide important new insights into the brain-wide dynamical changes underlying the psychedelic state - here provoked by LSD intake - and underline possible future clinical applications of psychedelic drugs in particular psychiatric disorders.

Keywords: Brain state; Functional MRI; LSD; Perturbation; Resting state networks; Whole-brain modelling.

Fig. 1

Calculation of the Perturbative Integration Latency Index (PILI). A. Initially, the computational whole-brain model was built based on the empirical structural connections between the 90 brain nodes. In this model each brain area was represented by a supercritical Hopf bifurcation. The model was fitted to the empirical functional connectivity in each of the 6 conditions, thus resulting in an optimal global coupling parameter for each condition. B. Next, we simulated the BOLD time series in each brain node for the basal dynamics and for the two perturbed states. The signals were band-pass filtered and Hilbert transformed to obtain the instantaneous phases and to subsequently calculate the phase locking matrix for each time point. C. Next, the integration was calculated as a function of time over 200 s in the basal state and after the offset of a model perturbation in either the synchronous or the noisy regime (here only shown the synchronous regime). The integration was computed by binarizing the phase locking matrix for different thresholds and calculating the number of areas in the largest connected component and finally integrating over thresholds. Finally the PILI was calculated, which characterizes the return of the brain dynamics to the basal state after a model perturbation of the system. For each trial, the PILI was computed as the integral under the curve of integration values after the offset of the model perturbation (yellow) until reaching the maximum of the basal state (blue). The final PILI was obtained by averaging over trials. (see Section 2. Methods for detailed explanation).

Fig. 2

Empirical functional connectivity and model fitting. In A the functional connectivity matrices are shown for each of the 6 conditions. Significance tests have been performed between the LSD and PCB conditions resulting in a significant difference in the mean functional connectivity between the LSD and the PCB state in the music scanning session. In B the mean and standard deviation over 50 realizations of the KS distance between the empirical and the simulated functional connectivity matrices are shown for each condition as a function of the global coupling strength. The optimal fit corresponds in each condition to the minimal KS distance. We found a significant difference between the optimal fit in the LSD and the PCB state in the music scanning session.

Fig. 3

Mean integration. The integration averaged over trials and nodes and the standard deviation of the integration over nodes is shown as a function of time for the three scanning conditions for both perturbation protocols. The mean and standard deviation of the integration are shown in dark green and light green for the basal state of the LSD and the PCB state, respectively. The mean and standard deviation of the integration are indicated in violet and orange and for the LSD and the PCB state, respectively.

Fig. 4

PILI - Node level analysis. Here the mean and the standard error of the mean (SEM) of the PILI values over trials are shown for each of the three scanning conditions for the LSD and the PCB state for all 90 brain regions. The vertical error bars represent the SEM for the PCB state and horizontal error bars represent the errors for the LSD state. The results show that the global differences between the LSD and PCB induced brain states were amplified in the music condition. Node-by-node analysis with corresponding p-values can be found in Table 1 and Supplementary Table S2.

Fig. 5

PILI - RSN analysis. The differences between the PILIs in LSD and PCB are shown on an RSN level. For all the nodes forming part of one RSN the Cohen's d value was calculated based on the mean and standard deviation over nodes in each state, indicating the standardized mean difference between the PILIs of each RSN in LSD and PCB. This was done for each of the 7 RSNs. The RSNs were ordered for each scanning condition (rest, rest with music, rest after music) by Cohen's d values, where darker colours indicate larger differences in PILI between the LSD and PCB conditions. The white area, which represents the corpus callosum and the subcortical structures, is to be discarded. It should be noted that the differences between PILI values in LSD and PCB state models for each RSN have found to be statistically significant in the rest and the rest with music condition. In the rest after music condition only 2 out of 7 networks (limbic network and DMN) show statistically significant differences (see Supplementary Table S4).

Fig. 6

Response variability. Here the distribution over trials of the standard deviation of PILI values is shown for the three different scanning conditions for LSD and PCB. Statistical differences between LSD and PCB brain states were evaluated with a two-sided t-test resulting in highly significant differences in all three scanning conditions with significantly higher PILI variability in the LSD state with respect to PCB. Especially in the music condition under the influence of LSD a considerably larger response variability can be observed with a p-value significantly smaller than 0.0001.