|
|
| Line 37: |
Line 37: |
| <sup>6</sup>Faculty of Social Science and Humanities, University of Ontario Institute of Technology, 2000 Simcoe Street North, Oshawa, ON L1H 7K4 Canada | | <sup>6</sup>Faculty of Social Science and Humanities, University of Ontario Institute of Technology, 2000 Simcoe Street North, Oshawa, ON L1H 7K4 Canada |
|
| |
|
| == Abstract==
| | {{:Store:Exel01}} |
|
| |
|
| An outstanding issue in cognitive neuroscience concerns how the brain is organized across different conditions. For instance, during the resting-state condition, the brain can be clustered into reliable and reproducible networks (e.g., sensory, default, executive networks). Interestingly, the same networks emerge during active conditions in response to various tasks. If similar patterns of neural activity have been found across diverse conditions, and therefore, different underlying processes and experiences of the environment, is the brain organized by a fundamental organizational principle? To test this, we applied mathematical formalisms borrowed from quantum mechanisms to model electroencephalogram (EEG) data. We uncovered a tendency for EEG signals to be localized in anterior regions of the brain during “rest”, and more uniformly distributed while engaged in a task (i.e., watching a movie). Moreover, we found analogous values to the Heisenberg uncertainty principle, suggesting a common underlying architecture of human brain activity in resting and task conditions. This underlying architecture manifests itself in the novel constant <math>K_{brain}</math>, which is extracted from the brain state with the least uncertainty. We would like to state that we are using the mathematics of quantum mechanics, but not claiming that the brain behaves as a quantum object.
| | {{:Store:Exel02}} |
|
| |
|
| | {{:Store:Exel03}} |
|
| |
|
| Subject terms: Computational science, Quantum mechanics
| | {{:Store:Exel04}} |
|
| |
|
| === Introduction ===
| | {{:Store:Exel05}} |
| An important but outstanding issue in contemporary cognitive neuroscience is understanding the organizational properties of neural activity. For instance, is there a fundamental structure to the spatial–temporal patterns neural brain activity across different conditions? One common approach used to address this question is to examine the brain at “rest”. Measures such as functional connectivity, independent component analysis and graph theoretic metrics, have been applied to data recorded using different imaging techniques (e.g., functional magnetic resonance imaging (fMRI) and electroencephalography (EEG)), to cluster brain areas that exhibit similar activity patterns. Numerous studies have shown that brain activity during “rest” can be grouped into distinct networks across<ref>Biswal, B., Zerrin Yetkin, F., Haughton, V. M. & Hyde, J. S. Functional connectivity in the motor cortex of resting human brain using echo-planar MRI. Magn. Reson. Med. 34, 537–541 (1995).</ref><ref>Hutchison,R.M.etal. Dynamic functional connectivity: Promise, issues, and interpretations. Neuroimage80,360–378(2013).</ref>; such as sensory (visual and auditory), default mode, executive, salience, and attentional (ventral and dorsal) networks that have been reliably reproduced across thousands of participants<ref>Eickhoff,S.B.,Yeo,B.T.T.&Genon,S.Imaging-basedparcellationsofthehumanbrain.Nat.Rev.Neurosci.19,672–686(2018).</ref>, and are predictive of phenotypic measures like cognition and clinical diagnoses<ref>Dajani DR, et al. Investigating functional brain network integrity using a traditional and novel categorical scheme for neurodevelopmental disorders. NeuroImage Clin. 2019;21:101678. doi: 10.1016/j.nicl.2019.101678. [PMC free article] [PubMed] [CrossRef] [Google Scholar]</ref><ref>Uddin LQ, Karlsgodt KH. Future directions for examination of brain networks in neurodevelopmental disorders. J. Clin. Child Adolesc. Psychol. 2018;47:483–497. doi: 10.1080/15374416.2018.1443461. [PMC free article] [PubMed] [CrossRef] [Google Scholar]</ref><ref>Sripada C, et al. Prediction of neurocognition in youth from resting state fMRI. Mol. Psychiatry. 2020;25:3413–3421. doi: 10.1038/s41380-019-0481-6. [PMC free article] [PubMed] [CrossRef] [Google Scholar]</ref>. These results suggest these networks may be an intrinsic aspect of neural activity.
| |
|
| |
|
| Indeed, the same set of structured patterns of neural activity have been found during "active" states, such as, while completing different tasks<ref>Biswal BB, Eldreth DA, Motes MA, Rypma B. Task-dependent individual differences in prefrontal connectivity. Cereb. Cortex. 2010;20:2188–2197. doi: 10.1093/cercor/bhp284. [PMC free article][PubMed] [CrossRef] [Google Scholar]</ref><ref>Fox MD, Raichle ME. Spontaneous fluctuations in brain activity observed with functional magnetic resonance imaging. Nat. Rev. Neurosci. 2007;8:700–711. doi: 10.1038/nrn2201. [PubMed] [CrossRef] [Google Scholar]</ref><ref>Kraus BT, et al. Network variants are similar between task and rest states. Neuroimage. 2021;229:117743. doi: 10.1016/j.neuroimage.2021.117743. [PMC free article][PubMed] [CrossRef] [Google Scholar]</ref>. For instance, there is a high degree of correspondence between networks extracted during rest and those extracted during tasks measuring sensorimotor<ref>Kristo G, et al. Task and task-free FMRI reproducibility comparison for motor network identification. Hum. Brain Mapp. 2014;35:340–352. doi: 10.1002/hbm.22180. [PMC free article][PubMed] [CrossRef] [Google Scholar]</ref><ref>Sui J, Adali T, Pearlson GD, Calhoun VD. An ICA-based method for the identification of optimal FMRI features and components using combined group-discriminative techniques. Neuroimage. 2009;46:73–86. doi: 10.1016/j.neuroimage.2009.01.026.[PMC free article] [PubMed] [CrossRef] [Google Scholar]</ref> and higher-level cognitive abilities (i.e., working memory)<ref>Calhoun VD, Kiehl KA, Pearlson GD. Modulation of temporally coherent brain networks estimated using ICA at rest and during cognitive tasks. Hum. Brain Mapp. 2008;29:828–838. doi: 10.1002/hbm.20581. [PMC free article] [PubMed] [CrossRef] [Google Scholar]</ref><ref>Xie H, et al. Whole-brain connectivity dynamics reflect both task-specific and individual-specific modulation: A multitask study. Neuroimage. 2018;180:495–504. doi: 10.1016/j.neuroimage.2017.05.050. [PMC free article] [PubMed] [CrossRef] [Google Scholar]</ref>.Even completing a task as complicated as following the plot of a movie elicits the same network architecture as observed in the resting brain<ref name=":0">Naci L, Cusack R, Anello M, Owen AM. A common neural code for similar conscious experiences in different individuals. Proc. Natl. Acad. Sci. U. S. A. 2014;111:14277–14282. doi: 10.1073/pnas.1407007111. [PMC free article] [PubMed] [CrossRef] [Google Scholar]</ref>. The correspondence between task and rest-based networks is so strong that task-based fMRI network activity can be predicted from the resting state<ref>Kannurpatti SS, Rypma B, Biswal BB. Prediction of task-related BOLD fMRI with amplitude signatures of resting-state fMRI. Front. Syst. Neurosci. 2012;6:7. doi: 10.3389/fnsys.2012.00007.[PMC free article] [PubMed] [CrossRef] [Google Scholar]</ref>, and rest-task network pairs can be identified at the individual level<ref>Elliott ML, et al. General functional connectivity: Shared features of resting-state and task fMRI drive reliable and heritable individual differences in functional brain networks. Neuroimage. 2019;189:516–532. doi: 10.1016/j.neuroimage.2019.01.068. [PMC free article][PubMed] [CrossRef] [Google Scholar]</ref>16. Together, these results suggest that rest and task-based patterns of brain activity likely share a similar underlying neural architecture, despite distinct experiences and cognitive processes<ref>Cole MW, Ito T, Cocuzza C, Sanchez-Romero R. The functional relevance of task-state functional connectivity. J. Neurosci. 2021 doi: 10.1523/JNEUROSCI.1713-20.2021. [PMC free article][PubMed] [CrossRef] [Google Scholar]</ref>.
| | {{:Store:Exel06}} |
| | |
| There are, however, important differences between the patterns of brain activity elicited during rest and task-based paradigms, and the set of experiences and cognitive processes associated with each<ref>Zhang S, et al. Characterizing and differentiating task-based and resting state fMRI signals via two-stage sparse representations. Brain Imaging Behav. 2016;10:21–32. doi: 10.1007/s11682-015-9359-7.[PMC free article] [PubMed] [CrossRef] [Google Scholar]</ref>. For instance, the presence or absence of a task is accompanied by increases in variability across different scales including neuronal firing rates changes in field potentials<ref>Monier C, Chavane F, Baudot P, Graham LJ, Frégnac Y. Orientation and direction selectivity of synaptic inputs in visual cortical neurons: A diversity of combinations produces spike tuning. Neuron. 2003;37:663–680. doi: 10.1016/S0896-6273(03)00064-3.[PubMed] [CrossRef] [Google Scholar]</ref><ref>Churchland MM, et al. Stimulus onset quenches neural variability: A widespread cortical phenomenon. Nat. Neurosci. 2010;13:369–378. doi: 10.1038/nn.2501. [PMC free article][PubMed] [CrossRef] [Google Scholar]</ref>, variation in fMRI blood oxygen level dependent (BOLD signal)<ref name=":1">He BJ. Spontaneous and task-evoked brain activity negatively interact. J. Neurosci. 2013;33:4672–4682. doi: 10.1523/JNEUROSCI.2922-12.2013. [PMC free article][PubMed] [CrossRef] [Google Scholar]</ref> and in EEG frequency bands<ref name=":2">Bonnard M, et al. Resting state brain dynamics and its transients: A combined TMS-EEG study. Sci. Rep. 2016;6:1–9. doi: 10.1038/srep31220. [PMC free article] [PubMed] [CrossRef] [Google Scholar]</ref> Furthermore, through transcranial direct current stimulation (tDCS) it has been shown that frontal-lobe stimulation increases one’s proclivity to mind wander <ref name=":8">Axelrod V, Zhu X, Qiu J. Transcranial stimulation of the frontal lobes increases propensity of mind-wandering without changing meta-awareness. Sci. Rep. 2018;8:1–14. doi: 10.1038/s41598-018-34098-z. [PMC free article] [PubMed] [CrossRef] [Google Scholar]</ref><ref name=":3">Axelrod V, Rees G, Lavidor M, Bar M. Increasing propensity to mind-wander with transcranial direct current stimulation. Proc. Natl. Acad. Sci. U. S. A. 2015;112:3314–3319. doi: 10.1073/pnas.1421435112. [PMC free article] [PubMed] [CrossRef] [Google Scholar]</ref>. Importantly, these differences are associated with changes in properties of neural activity but not in changes in the underlying neural architecture.
| |
| | |
| Is there a way to identify the shared neural architecture underlying the cognitive processes associated with rest and active states while also quantifying how these processes diverge from that shared architecture of neural activity? In this paper, we applied mathematical methods analogous to those of quantum mechanics, and the concept of phase space to EEG recorded during rest and movie-watching to extract spatial and transitional properties of dynamic neural activity. Quantum mechanics was developed to describe the dynamics of the subatomic world in terms of probability amplitudes and densities of states. Quantum systems (in the Schrodinger formulation of quantum mechanics) are described by wavefunctions which square to a probability distribution leading to the loss of local determinism and the Heisenberg uncertainty principle (for an overview/intro to the subject see<ref name=":7">Townsend JS. A Modern Approach to Quantum Mechanics.University Science Books; 2012. [Google Scholar]</ref>. This uncertainty principle places a fundamental limit on the location and the momentum of a point particle <ref>Heisenberg W. Über den anschaulichen Inhalt der quantentheoretischen Kinematik und Mechanik. Z. Phys. 1927;43:172–198. doi: 10.1007/BF01397280. [CrossRef] [Google Scholar]</ref>. In essence, if the position of a particle is known there is an underlying uncertainty in its momentum (one cannot precisely say how fast it is going) and vice versa. In addition to the adaptation of the wavefunction approach to quantum mechanics in this paper, we also employed a phase space model. Phase space is a widely used tool in the study of dynamical systems, where the positional variables are paired with their conjugate momenta which establishes a multidimensional space that describes all possible configurations of the given system. This space spans the entire range of states that a system can exist in, each point (in this hyper-space) represents a single state of the system. Phase space and its assorted formalisms are a classical concept, and we simply use it as another tool for analysing the EEG data. Herein, the mathematical methods of quantum mechanics are applied to EEG data to extract a proxy to phase space. This quasi-quantum approach naturally generates the concepts of ‘average’ position, ‘average’ momentum and culminates in an analogous Heisenberg uncertainty principle.
| |
| | |
| In this paper, we posit that using mathematical tools drawn from quantum mechanics, an underlying pattern representative of task and resting brain activity can be realised, in which differences across conditions are apparent, but culminates in a task independent constant value. It is important to note that we are not claiming that the brain behaves as a quantum object as some believe<ref>Penrose R. The Emperor’s New Mind. Viking Penguin; 1990. [Google Scholar]</ref> <ref>Penrose R. Shadows of the Mind: A Search for the Missing Science of Consciousness. Oxford University Press; 1994. [Google Scholar]
| |
| | |
| </ref> <ref>Atmanspacher H. Quantum Approaches to Consciousness.Stanford Encyclopedia of Philosophy; 2004. [Google Scholar]</ref><ref>Hameroff S. How quantum brain biology can rescue conscious free will. Front. Integr. Neurosci. 2012;6:93. doi: 10.3389/fnint.2012.00093. [PMC free article] [PubMed] [CrossRef] [Google Scholar]</ref>. Rather, we have employed some of the analytical tools from the Schrodinger formulation of quantum mechanics to the brain with the aim of gaining new insight into resting and task-based brain dynamics. Not only does devising this model probe questions into the functions of the brain, but it also provides a novel approach to analysing the myriad of data available in neuroscience.
| |
| | |
| === Results ===
| |
| In this paper, we adapted the probability amplitudes of quantum mechanics to define new metrics for examining EEG data—the ‘average position’ and ‘average momentum’ of the EEG signal. These were constructed from our definition of ‘brain states’ based on the quasi-quantum model. This allowed us to ascertain the frequency with which unique brain regions are entered by the pseudo-wavefunction, as well as explore the average-valued phase space. Finally, an analogous uncertainty relationship to that of quantum mechanics was established, with the full mathematical derivation described in the methods.
| |
| | |
| ==== Average values ====
| |
| The ‘average position’ of the EEG data was first extracted performing a Hilbert transform of the pre-processed time courses, and then applying a normalization constraint. Typically, the Hilbert transformed data is used to generate a metric of power dispersion or to extract the phase of the signal<ref>Freeman WJ, Vitiello G. Nonlinear brain dynamics as macroscopic manifestation of underlying many-body field dynamics. Phys. Life Rev. 2006;3:93–118. doi: 10.1016/j.plrev.2006.02.001.[CrossRef] [Google Scholar]
| |
| </ref><ref>le Van Quyen M, et al. Comparison of Hilbert transform and wavelet methods for the analysis of neuronal synchrony. J. Neurosci. Methods. 2001;111:83–98. doi: 10.1016/S0165-0270(01)00372-7.[PubMed] [CrossRef] [Google Scholar]</ref><ref>Freeman WJ. Deep analysis of perception through dynamic structures that emerge in cortical activity from self-regulated noise. Cogn. Neurodyn. 2009;3:105–116. doi: 10.1007/s11571-009-9075-3.[PMC free article] [PubMed] [CrossRef] [Google Scholar]</ref>. Instead, we imposed a new normalization condition, thereby creating an analogy to the wavefunctions of quantum mechanics. Denoting the Hilbert transformed time course of the <math>j</math>th electrode as <math>\Psi_j</math>, this is equivalent to
| |
| <center>
| |
| {| width="80%" | | |
| |-
| |
| | width="33%" |
| |
| | width="33%" | {{CD1}}<math>\Psi_j(t)=A_j(t)\exp (i\theta_j(t))</math>{{CD2}}
| |
| | width="33%" align="right" | <math>(1)</math>
| |
| |}
| |
| </center>
| |
| | |
| With <math>i=\sqrt{-1}</math>. We then imposed the normalization condition,
| |
| <center>
| |
| {| width="80%" |
| |
| |-
| |
| | width="33%" |
| |
| | width="33%" |{{CD1}}<math>\hat{\Psi}_j(t)= \frac{\Psi_j(t)}{\sqrt{\sum_{j=1}^{92}|\Psi_j|^2}}</math>{{CD2}}
| |
| | width="33%" align="right" |<math>(2)</math>
| |
| |}
| |
| </center>
| |
| | |
| | |
| | |
| The summation extends to 92, corresponding to the 92 electrodes selected from the original 129 on the head cap (channels removed from the face and neck for this analysis). This normalization constraint allowed us to define the probability at time <math>t</math> of the <math>j</math> ''j''th electrode as
| |
| <center>
| |
| {| width="80%" |
| |
| |-
| |
| | width="33%" |
| |
| | width="33%" | {{CD1}}<math>P_j(t)=\hat{\Psi}^*_j(t) \times\hat{\Psi}_j(t)</math>{{CD2}}
| |
| | width="33%" align="right" |<math>(3)</math><br />
| |
| |}
| |
| </center>
| |
| | |
| With the * denoting complex conjugation<ref name=":7" />. We then can describe each moment in time as a ‘brain state’ that is fully described (in the context of this model) through the ‘wavefunction’. This ‘brain state’ uniquely specifies the EEG signal, and hence the dynamics of interest, at each moment in time. Using this definition of probability, we defined two average quantities of interest. The average position and momentum are given explicitly by,
| |
| | |
| <center>
| |
| {|
| |
| | width="33%" |
| |
| | width="33%" |{{CD1}}<math>\langle x(t)\rangle=\sum_{j=1}^{92}x_jP_j(t)</math> <math>\langle p_x(t)\rangle=m{d \over dt}\langle x(t)\rangle=m\sum_{j=1}^{92}x_j{d \over dt}P_j(t)</math>{{CD1}}
| |
| | width="33%" align="right" |<math>(4)</math>
| |
| |}
| |
| </center>
| |
| | |
| | |
| | |
| With the same holding true for y. These two equations are how we create our quasi-quantum mechanical analogues. The second equation is an extension of Ehrenfest’s theorem, relating the average momenta of a particle to the time derivative of its average position. Where we have assumed a Hamiltonian with only a spatially dependent potential. Note that as the positions are fixed in space (positions of the electrodes) only the probability changes in time. Throughout this paper the mass m has been taking to be unity for both the <math>x</math> and <math>y</math> momenta. Each of the 92 electrodes were projected onto the horizontal plane, thus the <math>j</math>th electrode was described by one unique <math>(x_j,y_j)</math> point.
| |
| | |
| We first examined this model by grouping the 92 electrodes into eight regions on the scalp: Anterior L/R, Posterior L/R, Parietal L/R, Occipital L/R and the probabilities of each electrode in the region were summed to give a region-level probability. Figure 1A shows the <math>(x_j,y_j)</math> locations of each electrode, with different colours representing each of the eight groups. Figure 1B displays the frequency of entering each region, grouped by the four task conditions and two resting conditions. This reflects the normalized count of regional probabilities integrated in time. We found that each anterior region was entered more frequently while at rest than when subjects were engaged in either movie. Specifically, the anterior left and right regions had significant within stimulus change, with <math>P<0.001</math> (Tukey adjusted) for the ''Taken Rest—Taken, Taken Rest—Taken Scrambled, BYD Rest—BYD and BYD Rest—BYD Scrambled.'' This is in line with Axelrod and colleagues’ findings which showed activation in the frontal region was associated with mind wandering<ref name=":8" /><ref name=":3" />. We found frequency suppression in posterior regions, and an increase in anterior frequency in rest compared to the stimulated conditions, consistent with fMRI studies showing increased activation in the posterior cingulate cortex, and the medial prefrontal cortex during rest <ref name=":2" /><ref name=":3" /><ref name=":4">Wang RWY, Chang WL, Chuang SW, Liu IN. Posterior cingulate cortex can be a regulatory modulator of the default mode network in task-negative state. Sci. Rep. 2019;9:1–12. [PMC free article][PubMed] [Google Scholar]</ref><ref>Uddin LQ, Kelly AMC, Biswal BB, Castellanos FX, Milham MP. Functional connectivity of default mode network components: Correlation, anticorrelation, and causality. Hum. Brain Mapp. 2009;30:625–637. doi: 10.1002/hbm.20531. [PMC free article][PubMed] [CrossRef] [Google Scholar]</ref><ref>Stawarczyk D, Majerus S, Maquet P, D’Argembeau A. Neural correlates of ongoing conscious experience: Both task-unrelatedness and stimulus-independence are related to default network activity. PLoS One. 2011;6:e16997. doi: 10.1371/journal.pone.0016997.[PMC free article] [PubMed] [CrossRef] [Google Scholar]</ref><ref name=":5">Greicius, M. D., Krasnow, B., Reiss, A. L., Menon, V. & Raichle, M. E. ''Functional Connectivity in the Resting Brain: A Network Analysis of the Default Mode Hypothesis''. www.pnas.org. [PMC free article] [PubMed]</ref>. Thus, suggesting our model captures the frontal tendency associated with the brain activity while at rest.
| |
| | |
| [[File:Figure 1.jpeg|thumb|<small>'''Figure 1:'''</small> <small>('''A''') Electrode locations for each of the 92 electrodes on the Electrical Geodesics Inc. headcap. Electrodes were projected onto a horizontal plane with the nose in the positive y direction. Electrodes have been colour-coded to display the constituent parts of the 8 groups for the frequency analysis, namely, occipital left (blue)/right (orange), parietal left (green)/right (red), posterior left (purple)/right (brown) and anterior left (pink)/right (grey). ('''B''') Histograms representing the frequency of entering each region ''fG'' are displayed for the six conditions tested. Significant within stimulus change is present between each of the Anterior Left and Right regions when comparing the pre-stimulus rest and the respective stimulated condition (''P'' < 0.001, Tukey adjusted.). Error bars display the 1 standard deviation confidence interval.</small>|alt=|center|500x500px]]
| |
|
| |
|
| ===== Phase space ===== | | ===== Phase space ===== |