Brain Function: Novel Technologies Driving Novel Understanding - Bioinspired Approaches for Human-Centric Technologies - page 310

information channel is generated by stimulus-neural interactions (entrainment of
thalamic activity to the slow dynamics of naturalistic stimuli) and encodes infor-
mation about the temporal structure (rather than the strength) of slow stimulus
variations.
Advanced mathematical methodologies at IIT are also generating hypotheses
about neural circuit function that can be addressed by innovative new experimental
techniques, and reciprocally, new experimental approaches are demanding innova-
tive analytical approaches. For example, the ability to “causally” manipulate
neuronal circuits calls for better mathematical techniques to chart the flow of
information within the brain, going beyond classic correlative measures. Moreover,
specific predictions of computational models can be tested for the first time with
controlled activation of specific neuronal populations. For example, we are in a
position to test whether the low- and high-frequency independent information
channels are generated by different neuronal subtypes or by neuronal populations
in different cortical laminae. These fundamental questions will require an intimate
union of theory and experiment.

10.9 Multi-scale Neuroelectronic Brain Interfacing:
Challenges and Approaches

Current brain-interfacing technologies generally do not provide adequate spatial
and temporal resolution to access the activity of both single neurons and large
neuronal ensembles. Imaging techniques such as electroencephalography,
electrocorticography, magnetoencephalography, and functional magnetic reso-
nance provide real-time maps of the collective activity of large groups of neurons,
but all are coarsely limited in their temporal and/or spatial resolution. Optical
methods, described above, are a promising new method but likewise are still
restricted with respect to accessible spatial scale. Complementary approaches
must be generated to fill in the current spatiotemporal void in our understanding
of brain function.
Microelectrodes remain the most precise transducers of electrophysiological
signals from single neurons, with the resolution to detect spiking neural activity
(~1 kHz) and low-frequency field potentials (LFPs, 500 Hz). Classic microelec-

trodes allow recording from one neuron at a time, but modern multielectrode probes
allow recording from many neurons simultaneously. An archetypical
multielectrode probe includes a structured and implantable substrate to place
electrodes in the target brain areas as well as the electrical wiring that connects
each electrode to electronic circuits for signal conditioning, transmission, and
acquisition. These components advanced dramatically with the advent of microfab-
rication processes on silicon substrates in the 1970s. Current multielectrode probes
are used in a wide range of basic studies of brain function as well as for clinical and
neuroprosthetic applications. However, there is still a stringent need for increased

<