The name is absent



131

how the extra cost of solving the receptor ODEs and performing the reduction of the
voltage-dependent
Mg2+ block term impacts the speed of simulation, and how the
reduction of this term affects the accuracy of the system.

5.2 Augmenting the POD and DEIM Snapshot Sets

For cells with weakly excitable distal dendrites, the RSW model has shown the
promise of increasing accuracy when compared to the P0D+DEIM model alone, but
the accuracy of both can be improved by enriching the snapshot sets. Currently the
snapshot sets are generated from a single suprathreshold input, which is appropriate
for ion channel models with conductances that do not strongly attenuate the voltage
with distance from the soma. However, for models that yield weakly excitable den-
drites, the somatic spike may nearly disappear after it has propagated sufficiently far
from the soma. The result of this attenuation is that branches that are farther away
are underrepresented in the snapshot set, which yields a reduced system that is less
accurate than desired.

As mentioned in §3.6, a solution to this problem is to augment the snapshot set
with snapshots generated by multiple inputs to some of these more distal locations.
The inputs would need to be sufficiently strong to elicit the dynamics of interest,
otherwise the snapshots will contain too little information to appreciably improve
the accuracy of the model. The inputs should also be staggered so as to prevent
overlapping dynamics that would lead to a phenomenon similar to that of phantom



More intriguing information

1. On the job rotation problem
2. The name is absent
3. The name is absent
4. Disentangling the Sources of Pro-social Behavior in the Workplace: A Field Experiment
5. The name is absent
6. The Importance of Global Shocks for National Policymakers: Rising Challenges for Central Banks
7. Comparison of Optimal Control Solutions in a Labor Market Model
8. The name is absent
9. Olive Tree Farming in Jaen: Situation With the New Cap and Comparison With the Province Income Per Capita.
10. Types of Cost in Inductive Concept Learning