NeuroScience Neuroelectronic Interfacing: Semiconductor Chips with - - PowerPoint PPT Presentation

Nanoelectronics and Nanotechnology

NeuroScience

Neuroelectronic Interfacing: Semiconductor Chips with Ion Channels, Nerve Cells and Brain

João Abrantes nº 65693 Teresa Jorge nº 65722 Tomás Cruz nº 65725

NeuroScience

State of the art

Electrodes stimulation Brain Imaging and PET Scan

2 Nanoelectronics and Nanotechnology – 1st Semester 2011/2012 – MEFT

NeuroScience

State of the art

Zebrafish brain Optogenetics

3 Nanoelectronics and Nanotechnology – 1st Semester 2011/2012 – MEFT

Nanoelectronics and Nanotechnology – 1st Semester 2011/2012 – MEFT 4

Neurons

NeuroScience

Neuron Anatomy

Synapse Nucleous Body cell Dendrite Soma Axon Axon Terminal

5 Nanoelectronics and Nanotechnology – 1st Semester 2011/2012 – MEFT

NeuroScience

Different Neurons

6 Nanoelectronics and Nanotechnology – 1st Semester 2011/2012 – MEFT

NeuroScience

Synapse

Transfer of information between

neurons

Axon terminal. Synaptic vesicles. Neurotransmitters-biochemical

agents.

Synaptic cleft Receptors Dendrits

Nanoelectronics and Nanotechnology – 1st Semester 2011/2012 – MEFT 7

NeuroScience

Ion concentration

Lipid membrane separating interior from exterior medium.

Different concentration of Na+ and K+ ions inside and outside of the cell.

Higher concentration of K+ inside

Higher concentration of Na+ outside

8 Nanoelectronics and Nanotechnology – 1st Semester 2011/2012 – MEFT

NeuroScience

Diffusion potential

Diffusion occurs in the direction of the concentration gradient.

ions than to Na+ ions

9 Nanoelectronics and Nanotechnology – 1st Semester 2011/2012 – MEFT

NeuroScience

K+ and Na+ Pumps

To make the ions pass through the membrane against the concentration gradient the neuron has K+ and Na+ Pumps. This transport requires energy to happen. And happens at different rates for K+ and Na+.

10 Nanoelectronics and Nanotechnology – 1st Semester 2011/2012 – MEFT

NeuroScience

Action potential

 Transfer of information through

the neuron

Electrical signal

Propagates near the cell membrane.

Rest potential at -90mv

 3 phases

Rest Despolarization Repolarization How does it work?

11 Nanoelectronics and Nanotechnology – 1st Semester 2011/2012 – MEFT

NeuroScience Depolarization Membrane gets permeable to Na+, it enters the cell and increases voltage. Repolarization Membrane closes Na+ channels and opens K+ channel. Na+ stops flowing in, K+ flows out. Voltage decreases. Trigger is voltage.

Action potential

12 Nanoelectronics and Nanotechnology – 1st Semester 2011/2012 – MEFT

NeuroScience

Action potential

Depolarization Membrane gets permeable to Na+, it enters the cell and increases voltage. Repolarization Membrane closes Na+ channels and opens K+

• channel. Na+ stops flowing in, K+ flows out.

Voltage decreases. Trigger is voltage.

13 Nanoelectronics and Nanotechnology – 1st Semester 2011/2012 – MEFT

NeuroScience

Action potential

Beginning of action potential When the voltage rises a little above -90mv the action potential begins. Feedback The rising of the voltages opens the Na+

• channels. The opening of the channels allows

the voltage to rise faster. It is a feedback process. For the feedback process to suffice, the voltage should in the first impulse rise above a certain voltage. All or nothing Principle When there are conditions for the potential to propagate, it propagates. When there are not, it doesn't. There isn't half potentials.

14 Nanoelectronics and Nanotechnology – 1st Semester 2011/2012 – MEFT

NeuroScience

Action potential

Propagation of action potential The potential propagates to the neighbor membrane, in both directions.

15 Nanoelectronics and Nanotechnology – 1st Semester 2011/2012 – MEFT

Nanoelectronics and Nanotechnology – 1st Semester 2011/2012 – MEFT 16

Interface Model

NeuroScience

Micro/Nanoelectronic Devices Neurons

Hundreds of nanometers Micrometers Electrical signals Electrical signals

17 Nanoelectronics and Nanotechnology – 1st Semester 2011/2012 – MEFT

NeuroScience

À priori possible problems:

• Charge carriers are different and their mobility is very

different! Electrons in the electronic device and Ions in the neurons.

• Different architecture of the two information

processors.

Conclusion: Direct communication between single neuron and nanoelectronic device should be no big deal!

18 Nanoelectronics and Nanotechnology – 1st Semester 2011/2012 – MEFT

NeuroScience

Model of interface, what do we know already:

• Neuron is surrounded by a lipid membrane that is

a insulating material.

• Neuron membrane has proteins responsible for ionic

currents trough the membrane.

• Standard nanoelectronics, insulating layer over a

substratum.

19 Nanoelectronics and Nanotechnology – 1st Semester 2011/2012 – MEFT

NeuroScience

Most simple model, global contact. Membrane and insulator layer forms a compact dielectric:

• Easy to polarize and be polarized by the neuron.

This model doesn’t work, reality is just not that simple! (A variable voltage in the neuron directly polarizes the substratum and a variable voltage in the neuron)

20 Nanoelectronics and Nanotechnology – 1st Semester 2011/2012 – MEFT

NeuroScience

So, why does it not work? Big proteins and irregularities in the membrane don’t allow global contact. There is no compact dielectric.

21 Nanoelectronics and Nanotechnology – 1st Semester 2011/2012 – MEFT

NeuroScience

More realistic interface model: core coat conductor. A conductive cleft is created between the membrane and the insulator layer. The conductive cleft suppress mutual polarization and shield electric fields.

22 Nanoelectronics and Nanotechnology – 1st Semester 2011/2012 – MEFT

NeuroScience

Interface

• The current that flows across the cleft origins a

transductive potential between the membrane and the insulator layer.

• Polarization will occur but will be mediated by

the transductive potential. The current that flows in the cleft will be the base of the interface will have its origin in:

• Conductive current from the ion channels.
• Displacement current through the membrane.
• Displacement current through the insulator layer.

23 Nanoelectronics and Nanotechnology – 1st Semester 2011/2012 – MEFT

NeuroScience

Next simplest model: Point contact model.

Balance equation 1: Kirchhoff current law:

24 Nanoelectronics and Nanotechnology – 1st Semester 2011/2012 – MEFT

NeuroScience Balance Equation 2: Kirchhoff current law:

25 Nanoelectronics and Nanotechnology – 1st Semester 2011/2012 – MEFT

NeuroScience

Neuron-Silicon Circuits

Capacitive Stimulation Transistor Recording Neuronal Activity

26 Nanoelectronics and Nanotechnology – 1st Semester 2011/2012 – MEFT

Nanoelectronics and Nanotechnology – 1st Semester 2011/2012 – MEFT 27

Recording

NeuroScience

Transistor Recording of Neuronal Activity

• What will be the read of transistor ?

Membrane potential VM(t) Transductive Extracellular Potential VJ(t) Inside Neuron Cleft

28 Nanoelectronics and Nanotechnology – 1st Semester 2011/2012 – MEFT

NeuroScience

Transistor Recording of Neuronal Activity

• Calculation of VJ(t), according to Point-Contact Model

Balance Equations

Small Signal Aproximation

29 Nanoelectronics and Nanotechnology – 1st Semester 2011/2012 – MEFT

NeuroScience

Transistor Recording of Neuronal Activity

• Calculation of VJ(t), according to Point-Contact Model

Small Signal Aproximation (small VJ(t))

VE = 0

Current injected by a pipette

• VJ(t) is determined by the capacitive and ionic current through the attached membrane.
• VM(t) is governed by the currents through attached and free membrane.

30 Nanoelectronics and Nanotechnology – 1st Semester 2011/2012 – MEFT

NeuroScience

Transistor Recording of Neuronal Activity

• A-, B- and C-type response
• A-type response
• B-type response
• C-type response

1. The attached membrane contains no voltage-gated conductances (Vo

i=0);

2. Negligible leak conductance (gJM≈0).

฀ gJVJ  gJMVM  cM dVM dt VJ  dVM dt

1. The attached membrane contains no voltage-gated conductances (Vo

i=0);

2. Dominating ohmic leak conductance.

฀ gJVJ  gJMVM  cM dVM dt VJ VM

31 Nanoelectronics and Nanotechnology – 1st Semester 2011/2012 – MEFT

NeuroScience

Transistor Recording of Neuronal Activity

• Experiments of Neuronal Activity

Why neurons of invertebrates are preferred ?

• They are large and easy to handle;
• They form strong neuroelectronic junctions;
• Small networks may be reconstitued and studied on a chip.

Vs

32 Nanoelectronics and Nanotechnology – 1st Semester 2011/2012 – MEFT

NeuroScience

Transistor Recording of Neuronal Activity

• Experiments of Neuronal Activity – Leech Neuron

Cell Body of a neuron

• n the open gate oxide
• f field-efect transistor.

Axon stump of a neuron

• n a linear array of

field-efect transistors.

33 Nanoelectronics and Nanotechnology – 1st Semester 2011/2012 – MEFT

NeuroScience

Transistor Recording of Neuronal Activity

i. The cells are impaled with a micropipette and action potentials are elicited by current injection; ii. VM(t) is measured with the pipette; iii. The response of transistors is calibrated in terms of VJ(t) on the gate.

Experimental records

• Three types of records:

A – With a cell body on a transistor, B – With a cell body on a transistor, C – With a axon stump on a transistor,

฀ VJ(t) resembles dVM (t) dt ฀ VJ(t) resembles VM (t) ฀ VJ(t) resembles  dVM (t) dt

34 Nanoelectronics and Nanotechnology – 1st Semester 2011/2012 – MEFT

NeuroScience

Transistor Recording of Neuronal Activity

• Experiments of Neuronal Activity – Rat Neuron

Experimental conditions

i. Neurons from a rat hippocampus are cultures for seven days in neurobasal medium on a chip coated with polylysine; ii. Selected cells are contacted with a patch pipette;

• iii. Action Potentials are elicited by current pulses.

35 Nanoelectronics and Nanotechnology – 1st Semester 2011/2012 – MEFT

NeuroScience

Transistor Recording of Neuronal Activity

• Experiments of Neuronal Activity – Rat Neuron

Experimental records

Injection current

• Experimental Analysis
• VJ(t) ≈ 0,15 mV;
• VM(t) gives rise to two positive transients in VJ(t);
• The “jumps” in the record match the steps of

injection current;

• The small amplitude of the transistor record is a

consequence of a high conductance gJ wich is expected for the small size of rat neurons.

Potassium outward current Sodum inward current 36 Nanoelectronics and Nanotechnology – 1st Semester 2011/2012 – MEFT

NeuroScience

Transistor Recording of Neuronal Activity

• Final considerations

i. Neuronal Activity is detected by field-effect transistors; ii. The response is controled by VJ(t) in the cell-chip contact; iii. There is no only one response to action potentials, because the shape of VJ(t) depends on the cell type and the cell area attached to the chip; iv. The amplitude of recording signal is small, because the higher value of junction conductance gJ; v. The signals are weak for mammalian neurons due to their small size; vi. We may optimize recording by:

• Improving the cell-chip contact – reducing cleft width or improve specific

resistence;

• Enhancing the inhomogenety of channel distribution (membrane);
• Lowering the noise of transistors (improved design and fabrication).

37 Nanoelectronics and Nanotechnology – 1st Semester 2011/2012 – MEFT

Nanoelectronics and Nanotechnology – 1st Semester 2011/2012 – MEFT 38

Stimulation

NeuroScience

Next step in communication: Electronic Device Neuron According to the core coat condurtor model:

• A changing Vs leads to a displacement current through the

insulator layer.

• Current in the cleft appears and gives rise to a VJ.
• Voltage gated ions channels may open and a

action potential may arise.

39 Nanoelectronics and Nanotechnology – 1st Semester 2011/2012 – MEFT

NeuroScience

3 different types of response 3 different types of stimulation

Type A stimulation: ion conductance neglectable.

In this approximation, balance equation 2 leads to: And balance equation 1:

Solving this to a step stimulation the result will be:

with Response time of ion gated cells should be bigger. Nothing happen??

40 Nanoelectronics and Nanotechnology – 1st Semester 2011/2012 – MEFT

NeuroScience

Type B stimulation: ion conductance is important, VJ >> VM

With this approximation, balance equation 1 leads to: That can be solved for a step stimulation:

with

After some hard and boring calculations that can be obtained: Now the voltage gated ion channels will see a potential:

41 Nanoelectronics and Nanotechnology – 1st Semester 2011/2012 – MEFT

NeuroScience

If this VM is over some certain value of threshold a action potential may arise.

Experiment with leech neurons, stimulated by step potential of different magnitudes:

• The action potential as a ms time

scale

• An action potential arises

when VM reaches the threshold value.

• Depending on VS0, the action

potential is achieved faster or slower.

42 Nanoelectronics and Nanotechnology – 1st Semester 2011/2012 – MEFT

NeuroScience

In type C stimulations we don’t really kwon what is going on because there could be very different responses.

In this experiment can be seen all that we have said, we excite neuron with an applied VS, read VM with a micropipette and read VJ with a transistor. The results are:

43 Nanoelectronics and Nanotechnology – 1st Semester 2011/2012 – MEFT

NeuroScience

• The immediate response to

burst voltages applied are short capacitive transients.

• After third pulse, the

threshold is achieved and an action potential arise.

44 Nanoelectronics and Nanotechnology – 1st Semester 2011/2012 – MEFT

Nanoelectronics and Nanotechnology – 1st Semester 2011/2012 – MEFT 45

Neuron Nets

NeuroScience

Neuronal Nets

• The functions of neuronal networks:
• The requirements of experiments on netwok dynamics:

i. Definition of the growth area of a set of neurons; ii. Enhancement of synaptic strenght as a consequence of correlated presynaptic and postsynaptic activity.

i. A noninvasive supervision of all neurons with respect to stimulation and recording to observe the performance of the net on a long time scale; ii. A fabrication of neuronal maps with a defined topology of the synaptic

• connections. For this, we have to control neuronal outgrowth.

46 Nanoelectronics and Nanotechnology – 1st Semester 2011/2012 – MEFT

NeuroScience

Neuronal Nets

• How we control the neuronal outgrowth ?

 Chemical guidance;  Topographical guidance;  Electrical guidance.

47 Nanoelectronics and Nanotechnology – 1st Semester 2011/2012 – MEFT

NeuroScience

Neuronal Nets

• Chemical Guidance

The motion of neuronal growth cones is guided by chemical

• guidance. Defined arborizations of leech neurons are achieved by

chemical guidance with lanes of extracellular matrix protein wich are fabricated by UV photolithography. Using linear chemical patterns, we are able to guide the outgrouwth of a pair of

• neurons. Their growth cones are forced to collide and to form a synaptic connection.

Experiment of a leech neuron Experiment of snail neurons

48 Nanoelectronics and Nanotechnology – 1st Semester 2011/2012 – MEFT

NeuroScience

Neuronal Nets

• Chemical Guidance
• Problems

1. The stabilitic of neuritic trees is limited, because neurites have tendency to shorten; 2. The substrate provides a restricted area of growth, but it doesn´t define a certain direction of neurites.z

Electronmicrograph of the leech neurites

The shape of a neuritic tree is lost and neuritic tree is not defined by the pattern !

49 Nanoelectronics and Nanotechnology – 1st Semester 2011/2012 – MEFT

NeuroScience

Neuronal Nets

• Topographical guidance

The instability of grown neuritic trees in chemical guidance is

• vercome. There the grown neuritics are immobilized by

microscopic grooves. The synaptic connections are checked by impaling with micropipettes.

• Problem

Network of three snail neurons

A neuritic tree is not uniquely defined by the guiding pattern !

• Electrical guidance

In order to solve above problems, we may use electrical manipulation at certain positions and times to promote or inhibit the growth and induce or prevent synapse formation.

50 Nanoelectronics and Nanotechnology – 1st Semester 2011/2012 – MEFT

NeuroScience

Neuronal Nets

• Final considerations

The use of previous methods is still uncertain. An alternative strategy may undergo uncontrolled growth of neurons on a chip consisting of thousands of closed structures of communication. In this sense, the neurons could be in contact at any time and rearrange a network that would be supervised in its structure and dynamics.

51 Nanoelectronics and Nanotechnology – 1st Semester 2011/2012 – MEFT

Nanoelectronics and Nanotechnology – 1st Semester 2011/2012 – MEFT 52

Brain Slices

NeuroScience

Brain Silicon Chips

Brain Slices (For neuron net study) Advantages Already grown neuronal nets. Few layers cell thick Conserve major neuronal connections. Disadvantages Dead tissue. Preference for planar networks due to planar nature of silicon chips. Connection with single neuron impossible. Stimulation and recording of populations of neurons

53 Nanoelectronics and Nanotechnology – 1st Semester 2011/2012 – MEFT

NeuroScience

Brain Silicon Chips Model

There are two models for the brain silicon chips. The first we discuss is the Volume Conductor, where you consider the brain slice to have volume. This is not the most used model.

54 Nanoelectronics and Nanotechnology – 1st Semester 2011/2012 – MEFT

NeuroScience

Sheet-conductor model

In this model we consider the brain tissue to be 2D.

Ohmic conductance “shutting effect of the bath” Resistance “thin tissue” Capacitance “substrate” Neuronal current source Stimulation current Due to changing voltage Vs in the substrate Continuity/ Conservation relation of current

55 Nanoelectronics and Nanotechnology – 1st Semester 2011/2012 – MEFT

NeuroScience

We will analyze a series of stimulated and recorded signals across a rat hippocampus.

Transistor recording of brain slice

Gyrus dentatus

The stimulation occured in the DG area and the recording was performed in the CA1 area. In two deiferent layers: Stratum radiatum (dendrites) Stratum pyramidale (cell bodies)

56 Nanoelectronics and Nanotechnology – 1st Semester 2011/2012 – MEFT

NeuroScience In the pic you can see: Field potentials in the stratum radiatum (top) and in stratum pyramidale (bottom)

• f the CA1 region.

The arrows mark the stimulation artifact The dashed line stands for the measurement with a micropipette electrode. The transistor records with higher voltage resolution. In the dendrites the signal in negative, potential going in. In the cell body is positive, potential coming out.

Transistor recording of brain slice

57 Nanoelectronics and Nanotechnology – 1st Semester 2011/2012 – MEFT

NeuroScience

Model simplification for recording

1D approach to the hippocampus slice. No stimulation Neglecting the capacitive current . The current profile considered and the one physiologically meaningful is

• ne with constant density of inward current density in the dendrites

and balanced constant outward current in the body cells.

58 Nanoelectronics and Nanotechnology – 1st Semester 2011/2012 – MEFT

NeuroScience

Model simplification validation

Profile of field potential across the CA1 area. Black Dots - Amplitude of voltage transient plotted versus the position of the transistor. (Experimental) The data are fitted with the Vfield computed using the sheet-conductor model with the presented modifications and with profile equal to the dashed line. The thin lines are for the potential in the stratum pyramidale and radiatrum. The bulk line is the sum of the other two.

59 Nanoelectronics and Nanotechnology – 1st Semester 2011/2012 – MEFT

NeuroScience

Capacitive Stimulation

We will analyze the stimulation for a specific potential. In this case it will be a circular potential. So we use cylindrical coordinates. And assume Stationary stimulation implies and

60 Nanoelectronics and Nanotechnology – 1st Semester 2011/2012 – MEFT

NeuroScience

Capacitive Stimulation

In this case we don’t have any experimental data. We evaluate through the sheet- conductor model which would be the applied field to a slice in case of Being according the dashed line. The amplitude of the stimuli expected in the neuron is much smaller than the

• ne given to the cell.

The bulk line is the sum of the two thin lines. We can make the stimuli more precise by summing a negative potential, with the specified geometry, to a positive one. But this further decreases the amplitude.

61 Nanoelectronics and Nanotechnology – 1st Semester 2011/2012 – MEFT

Nanoelectronics and Nanotechnology – 1st Semester 2011/2012 – MEFT 62

Bibliography

NeuroScience

• Fromherz, Peter. Neuroelectronic Interfacing: Semiconductor Chips with Ion Channels,

Nerve Cells, and Brain. Berlim, 2003.

• Arthur C. Guyton, MD, and Jonh E. Hall, PhD. Tratado de Fisiologia Médica, translation
• f 11th edition
• Barbara Young, James S.Lowe, Alan Stevens, Jonh W. Heath. Histologia Funcional,

translation of 5th edition, Wheater.

NeuroScience

The End

63 Nanoelectronics and Nanotechnology – 1st Semester 2011/2012 – MEFT