Functional Connectivity of In-Vitro Neuronal Spiking Activity during Rest and Gameplay
Poster No:
1171
Submission Type:
Abstract Submission
Authors:
Forough Habibollahi1, Moein Khajehnejad2, Alon Loeffler1, Adeel Razi2,3, Brett Kagan1
Institutions:
1Cortical Labs, Melbourne, VIC, 2Turner Institute for Brain and Mental Health, Monash University, Melbourne, VIC, 3Wellcome Centre for Human Neuroimaging, University College London, United Kingdom
First Author:
Co-Author(s):
Adeel Razi
Turner Institute for Brain and Mental Health, Monash University|Wellcome Centre for Human Neuroimaging, University College London
Melbourne, VIC|United Kingdom
Turner Institute for Brain and Mental Health, Monash University|Wellcome Centre for Human Neuroimaging, University College London
Melbourne, VIC|United Kingdom
Introduction:
The DishBrain system combines biological intelligence with adaptive neuronal traits by seamlessly integrating in-vitro neuronal networks with in silico computational elements using high-density multi-electrode arrays (HD-MEAs) (Kagan et al, 2022). These cultivated neuronal ensembles exhibit biologically-based adaptive intelligence, replicated effectively in a dynamic gaming environment through closed-loop stimulation and concurrent recordings. These neuronal ensembles demonstrate self-organized adaptive electrophysiological dynamics, responding meaningfully to biologically plausible external stimuli. Empirical data is sourced from cortical cells derived from embryonic rodent or human induced pluripotent stem cell lineages. While synthetic biology methods show that in-vitro neuronal networks can achieve real-time adaptive learning in simulated environments, underlying network dynamics associated with this learning remain unexplored.
Methods:
The DishBrain system simulated neural cultures in a virtual 'Pong' game. Stimulation involved rate coding pulses (4Hz to 40Hz) for the ball's x-axis and place coding (specific electrodes) for the y-axis within an 8-electrode sensory zone. Paddle movement corresponded to real-time electrophysiological activity in a 'motor area', with cultures receiving feedback on paddle movement. Missed balls triggered a 150 mV, 5 Hz unpredictable stimulus on random sensory electrodes for 4 seconds, followed by a 4 second resting period. Each gameplay session lasted 20 minutes at a 20 kHz sampling rate. Neuronal spiking was captured from 900 HD-MEA channels across 248 Gameplay and 147 Rest sessions. First, t-SNE algorithm (Van der Maaten, 2008) was utilized to generate 3D lower-dimensional representations of these lengthy spiking trains. To further reduce computational complexity, a method to identify a subset of recorded channels likely monitoring neuronal populations relevant to the task was developed. Employing Tucker decomposition through higher-order orthogonal iteration on tensor data from all Gameplay sessions in the lower-dimensional embedding space, we derived a 900 × 3 tensor, capturing underlying patterns. Using this tensor, we applied the K-medoid clustering algorithm to identify representative channels as 'medoids' of the created 30 clusters. Subsequently, a network matrix was constructed using functional connectivity, defined as zero-lag Pearson correlations, between these 30 channels as nodes. Only edges with Pearson correlation absolute values exceeding 0.7 were retained.
Results:
To analyze connectivity network evolution, we segmented each recording session into 2-minute windows and tracked changes in edge weights over time. Cultures in the game environment exhibited a higher number of edges with increased correlation between channels, unlike during rest state spontaneous activity. We assessed various macroscopic network metrics such as average weight, modularity index, clustering coefficient, max betweenness, and characteristic path length. Comparisons were made between Rest and Gameplay states, as well as the initial and final 2 minutes of recordings in each group. All of these metrics except the characteristic path length showed statistically significant differences using one-way ANOVA during Gameplay (p = 2.265e-3, 8.478e-8, 1.891e-6, 1.005e-4, and 0.071, respectively), but not in the Rest condition of the cultures (p = 0.864, 0.670, 0.738, 0.281, and 0.899, respectively).
Conclusions:
Our findings offer insights into the dynamic learning traits of biological cells within a structured information landscape. Our innovative approach facilitates:
1) studying neurons at a cellular level in a closed-loop game environment, 2) analyzing neuronal activities in a lower-dimensional space, 3) identifying an efficient subpopulation capturing overall dynamics, and 4) examining network dynamics influencing information processing and learning.
1) studying neurons at a cellular level in a closed-loop game environment, 2) analyzing neuronal activities in a lower-dimensional space, 3) identifying an efficient subpopulation capturing overall dynamics, and 4) examining network dynamics influencing information processing and learning.
Brain Stimulation:
Direct Electrical/Optogenetic Stimulation
Learning and Memory:
Neural Plasticity and Recovery of Function
Learning and Memory Other 2
Modeling and Analysis Methods:
Connectivity (eg. functional, effective, structural) 1
Keywords:
Computational Neuroscience
ELECTROPHYSIOLOGY
Learning
Neuron
Plasticity
1|2Indicates the priority used for review
Abstract Information
By submitting your proposal, you grant permission for the Organization for Human Brain Mapping (OHBM) to distribute your work in any format, including video, audio print and electronic text through OHBM OnDemand, social media channels, the OHBM website, or other electronic publications and media.
The Open Science Special Interest Group (OSSIG) is introducing a reproducibility challenge for OHBM 2025. This new initiative aims to enhance the reproducibility of scientific results and foster collaborations between labs. Teams will consist of a “source” party and a “reproducing” party, and will be evaluated on the success of their replication, the openness of the source work, and additional deliverables. Click here for more information. Propose your OHBM abstract(s) as source work for future OHBM meetings by selecting one of the following options:
Please indicate below if your study was a "resting state" or "task-activation” study.
Healthy subjects only or patients (note that patient studies may also involve healthy subjects):
Was this research conducted in the United States?
Were any human subjects research approved by the relevant Institutional Review Board or ethics panel? NOTE: Any human subjects studies without IRB approval will be automatically rejected.
Were any animal research approved by the relevant IACUC or other animal research panel? NOTE: Any animal studies without IACUC approval will be automatically rejected.
Please indicate which methods were used in your research:
For human MRI, what field strength scanner do you use?
Which processing packages did you use for your study?
Provide references using APA citation style.
1. Kagan, B. J. et al. (2022). In vitro neurons learn and exhibit sentience when embodied in a simulated game-world. Neuron, 110(23), 3952-3969.
2. Van der Maaten, L. et al. (2008). Visualizing data using t-SNE. Journal of machine learning research, 9(11).