Micro-Electrode Arrays technology and

in vitro neuronal models

Roberto Raiteri

Department of Informatics, Bioengineering, Robotics and System Engineering

University of Genova

UTB Cartagena february 19 2020

The brain

• weight:  1.5 kg (2% of body weight)

•  1011 neurons + glia cells (astrocytes,
microglia, oligodendrocytes)

UNIGE @ ILMENAU
total surface area of the cerebral cortex = 0.25
th
m2
12 December 2012
 The neuron [1/4]
The neuron is the functional unit of the
nervous system. It is made up of:
• soma (cell body)
• dendrites (which generally receive input from
other neurons)
• axon (to generate an output)
• pre-synaptic and post-synaptic regions (to allow
the communication among neurons)

UNIGE @ ILMENAU
12th December 2012
 The neuron [2/4]
• An Action potential is a fast transient
variation of the (trans)membrane
potential which propagates along the
neuronal membrane of the excitable cells
• There are no big or small action
potentials in one neuron - all action
potentials have the same amplitude.
Therefore, the neuron either does not
reach the threshold to activate an action
potential or a full action potential is fired
("ALL OR NONE" principle)

In 1963 Hodgkin and Huxley got the Nobel Prize

for medicine because their first who recording of
an action potential and the development of a
mathematical model of such process UNIGE @ ILMENAU
12th December 2012
The neuron [3/4]
Synapses are cell structures which allow a neuron (presynaptic) to pass an
electrical signal like an action potential to another neuron (postsynaptic) or
to the target effector cell (e.g., muscle)

excitatory
Possible classification of synapses:
morphology: physiology:
inhibitory
- chemical
- electrical (gap junction)

UNIGE @ ILMENAU
12th December 2012
The neuron [4/4]
Synapses are essential to neuronal function, and are at the basis of the
brain function (in the cortex 2*1010 neurons and 1.5*1014 synapses)

UNIGE @ ILMENAU
12th December 2012
 Rationale
The capability of the brain to encode and
process incoming stimuli is affected by
both structure and ongoing dynamics

The brain’s structural and functional

connectivity exhibits peculiar features of
complex network topologies (neither (Sporns, Ann. N.Y. Acad. Sci.,
completely connected with each other nor 2011)

randomly linked; instead, their

interconnections show a specific and
intricate organization)

There is a need for reduced and

engineered model systems to address the
above issues
(Marguet and Harris, UNIGE
J. Neurosci., 2011)
@ ILMENAU
12th December 2012
Reducing the complexity … towards
brain-on-a-chip

degree of anatomical complexity

20 s

electrophysiological dynamics modulation

UNIGE @ ILMENAU
12th December 2012
 Outline
• Micro Electrode Arrays

• In vitro neuronal networks

• Recording and stimulation

with MEA

• Engineered neuronal
networks: from 2D to 3D UNIGE @ ILMENAU
12th December 2012
 Outline
• Micro Electrode Arrays

• In vitro neuronal networks

• Recording and stimulation

with MEA

• Engineered neuronal
networks: from 2D to 3D UNIGE @ ILMENAU
12th December 2012
 Micro-Electrode Arrays (MEAs) on
the market
devices for long-term extra-cellular recording and stimulation of
the electrophysiological activity of neuronal assemplies
fabrication technology: thin-film technology (micro- and nano-
fabrication)
devices on the market:

Ayanda-Biosystems Multichannel Systems (MCS) Panasonic

UNIGE @ ILMENAU
12th December 2012
MEAs: an enabling technology to complement
traditional electrophysiological techniques and
biosensing applications
conventional features
material insulating glass
 long-term culturing (from days
number of electrodes 60 - 120
to months)
electrode diameter 30 µm
 multi-site extracellular
pitch 100 µm
recording / stimulation
acquisition type real-time
 study / design of in-vitro sampling rate 10-20 kHz
experimental models

UNIGE @ ILMENAU
12th December 2012
MEA fabrication process
A-A

A A

1. pyrex substrate (or silicon)

2. photoresist coating and photolithography

3. metal evaporation (Pt, Ir, Au)

4. lift-off
MEA fabrication process
A-A

A A

5. insulation - silicon nitride

6. photoresist coating and photolithography

7. plasma etching
MEA fabrication process
A-A

A A

5. insulation - silicon nitride

6. photoresist coating and photolithography

7. plasma etching

8. photoresist stripping and cleaning

Example: MEAs electrode materials and
morphologies
planar
electrode materials 10 m
Au, Pt, ITO, IrOx, TiN

hillocks

20 m

tips

47 m
Example: MEA with a clustering structure

• microelectrode array
Pyrex 7740 substrate (14 x 14 mm2)
60 Pt electrodes
electrodes diameter: 30 μm
• clustering structure
EPON SU-8 with adhesion layer
346 ± 8 μm thick
wells diameter: 3 mm
channels: 300 μm x 800 μm
• temperature sensors
Pt-RTDs of 1 kΩ

UNIGE @ ILMENAU
12th December 2012
Thin-film technology
key features
- accurate control of size, shape (micron) and thickness (nano)

- parallel (batch) processing (at the wafer level)

- “wide” range materials:

- standard SiO2, Si3N4, polySi, Al
- non standard
metals Pt, Au, Ag, Ir, C
oxides TiO2, Al2O5, Ta2O5
nitrides TiN
polymers SU-8, PDMS, PI

- “flexible” technology
- organic materials (organic transistors→flexible devices)
UNIGE @ ILMENAU
12th December 2012
High-density CMOS MEA devices
CMOS technology: each recording «electrode» is a transistor with
an integrated pre-amplifier

no. of electrodes 4096 electrode/pixel

electrode size 21 µm
electrode separation 21 µm
active area ~ 7 mm2
spatial density ~ 580 el/mm2
sampling rate 7.7 - 125 kHz
data rate ~ 0.5 Gbit/s

High-resolution Low-resolution
(~580 el/mm2) ~19 el/mm2

Large-scale Covered area:

(~ 7 mm2) ~2,7 mm2

UNIGE @ ILMENAU
12th December 2012
Summary on MEA devices

Different materials, different shapes: thin film technology

Main characteristics:

• non invasive techniques (extracellular measurement)

• recording and stimulation
• transparent substrate (not for CMOS MEAs)
• re-usable easy-to-use

60-4096 electrodes
7-50 µm electrode dimension
10-500 µm electrode pitch
 Outline

• Micro Electrode Arrays

• In vitro neuronal networks

• Recording and stimulation

with MEA

• Engineered neuronal
networks: from 2D to 3D
UNIGE @ ILMENAU
12th December 2012
 in vitro neuronal networks
Dissection
Cultured networks From rat-mouse embryos +
(or newborn) Enzymatic digestion
+
Mechanical dissociation

50μm

20μm
 pretty different from the in vivo architecture, but some fundamental
mechanisms are maintained.
 main receptors for neurotransmitters of the CNS (i.e. glutamate and GABA) are
expressed after two weeks in culture (Köller et al., 1990). UNIGE @ ILMENAU
12th December 2012
Other in vitro experimental models [1/2]
Brain slice (acute or organotypic)

 There is a good fidelity with the in

vivo architecture
 They can only survive for short
period (i.e. hours – acute slices)
 They show little spontaneous
activity

Intact brain: explanted brain

 A real brain !
 Only connection to the
periphery and to the body
are missing
 It can survive only few hours
UNIGE @ ILMENAU
12th December 2012
Other in vitro experimental models [2/2]

Brain organoid
Typically from human –
induced Pluripotent Stem Cells
(h-iPSC)

Model of the developing brain,

used as in vitro neural model
of neurodevelopmental
disorders

UNIGE @ ILMENAU
12th December 2012
 Outline

• Micro Electrode Arrays

• In vitro neuronal networks

• Recording and stimulation

with MEA

• Engineered neuronal
networks: from 2D to 3D
UNIGE @ ILMENAU
12th December 2012
In vitro neuronal networks onto MEAs
Primary cultures of rat cortical neurons

Dissection
+
Enzymatic digestion
+
Mechanical dissociation

~ 50000 cells
Rat embryos (E18)

20 V
100 ms
10 ms Spike

5 cm 1.64 mm

30μm

20 days
Extracellular field potential recording

Only changes in the «membrane voltage» are measured:

 action potentials, local field
potentials, …
 no resting state

Electrogenic cell

Cell-electrode junction
Metal electrode
Intra- vs. extra-cellular recording
mV

Culture media
Electrogenic cell

intracellular recording: V12

1
3 (RE)

2 (WE)
Metal electrode

extracellular recording: V23

action potential: V12 = up to 100 mVp-p V23 = Ve = 20-200 Vp-p

(in rat’s cortical neurons)
Electrophysiological signals
Primitive patterns of synchronized electrical activity are observed, usually
ranging from stochastic “spiking” to organized “bursting”

› Bursting is a peculiar behavior

shown by developing in vitro
neuronal systems. It
represents an exploring
dynamics of the in vitro
network which has to find a
stable state in which synaptic
connections are properly
formed.
› This activity pattern changes
its features along with the
maturation level of the
network.

UNIGE @ ILMENAU
12th December 2012
Spatially resolved MEA recording
 Computational neuroscience: the interplay
between dynamics and connectivity in neuronal
networks

Inferring functional connectivity and

topological properties of networks based on
• cross correlation excitatory

map of functional links

• transfer entropy

inhibitory
 In vitro neuronal networks onto MEAs

Not only recording but also electrical stimulation

Voltage or current stimulation

Post
Stimulus
Time
Histogram
Summary on NN and MEA recordings
• dissociated neurons can be grown onto MEA substrates
(molecules that promote cell adhesion are needed)

• coupling with slices and organoids

• both network and “single neuron recording”

electrophysiology recordings

• long-term recordings: 1-2 months standard; up to 3-4

months

• easy to electrically stimulate

• easy to couple with other stimulation methods: e.g.

optogenetics (sonogenetics)
 Outline

• Micro Electrode Arrays

• In vitro neuronal networks

• Recording and stimulation

with MEA

• Engineered neuronal
networks: from 2D to 3D
UNIGE @ ILMENAU
12th December 2012
Engineered 2D networks

patterning adhesion promoters / inhibitors

Ink-jet printing

patterning by mechanical constraints

• PDMS channels
• neurocages
• photo-thermally-etched agarose

UNIGE @ ILMENAU
12th December 2012
Case study: PDMS structures and dual chambers

Standard MEA 200/30 Micro-channels features

TiN/SiN w/o ring with width 10 m
60 microelectrodes
arranged in a 8 x 8 length 150 m
layout Micro-fluidic
compartment height 3 m
width 1.5 mm channel
50 m
spacing
length 8 mm
#
height 100 m 120
channels
reservoir Ø 6 mm

Polydimethylsiloxane - PDMS Kanagasabapathi T., et al.,. Frontiers in Neuroengineering, Vol. 4, 2011

 Guided neurites growth

Cell bodies remain

confined in the relative
compartments

Only bundle of neurites go

through the microchannels

• Segregation and possibility to recreate

interconnected sub-populations
• Heterogeneous networks (co-cultures):
cortex-thalamus UNIGE @ ILMENAU
cortex-striatum 12th December 2012
 Interacting populations

Kanagasabapathi T., Massobrio P., et al., Functional connectivity and dynamics of cortical-thalamic UNIGE @ ILMENAU
networks co-cultured in a dual compartment device. Journal of Neural Engineering, Vol. 9, N.3, 2012. 12th December 2012
Engineered 3D networks
Structure developed in a 3D space onto a 3D scaffold (vs. organoids)
Motivation: environment, neuronal morphology and interaction with
glial cells and extracellular matrix more similar to in vivo situation

An example: 3D neuronal network based on self assembled glass microbeads

UNIGE @ ILMENAU
12th December 2012
Engineered 3D networks

50 m MAP-2 green
Neu-N: red

 neurons fill the interstitial space of the self-

assembled glass microbeads
MAP-2: red
 neuronal soma have a round shape
20 m
 neurites develop in a 3D space and form@extensive
UNIGE ILMENAU
Frega et al. Scientific Report, (2014) arborizations connecting the neurons 12th December 2012
 Electrophysiology: 2D, 3D and controls

UNIGE @ ILMENAU
15 s 15 s 12th December 2012
 Electrophysiology: 2D vs 3D
Registration of 30 min

2D Sites where the bursts started 3D

511 173 network bursts
network bursts

Few sources with high occurrence A lot of sources with low occurrence

UNIGE @ ILMENAU
12th December 2012
Chitosan based scaffolds for 3D neuronal
Biopolymer networks
Chitosan based scaffolds for 3D neuronal
networks
 Summary on 2D and 3D engineered
neuronal cultures
• patterned 2D cultured networks onto MEA
• chemically
• physically (e.g. microfluidics made with elastomer coupled with
MEA)

• 3D cultures onto scaffolds and coupled to MEA devices

for network electrophysiology

• Chitosan as biomaterial for advanced and bio-integrable

scaffolding for neuronal cells

• Open issue: 3D electrodes for 3D recording/stimulations

 Summary and conclusions

• State-of-the-art technologies allow for in vitro human based neuronal

models (human induced-Pluripotent Stem (h-iPS) cells) that can be
studied in terms of network electrophysiology

• This allows to investigate the capability of neural systems to encode

and process stimuli and the mechanisms of computation, i.e. brain
functions and dysfunctions

• In vitro 3D models towards engineered brain-on-a-chip applications:

patient specific medicine

• Engineered model systems are needed to address issues in complex

networks: e.g., dynamics and connectivity, interactions among
populations, representation of information
The Polytechnical School (School of Engineering) of the University of
Genova offers a two-year degree «Luarea Magistrale» in Bioengineering,
with a specific curriculum in «Neuroengineering and neurotechnologies»