Title: Neural Recording System
1Neural Recording System
- By
- Ivana Duvnjak
- Stephen Kilpatrick
- Dawit Fsiha
2Outline
- What is a Neural Network Recording System?
- Why are they important?
- How can they be improved?
- Some possible solutions
- Analysis of some solutions
- Conclusions
3Neural Network Recoding Systems
- Brain sends electrical signals - Neurons
- Sent through axons
- Electrical Signals processed by synapses
- Mapping the brain
4Applications
- Medical Research
- Mental Performance
- Mental Illness
- Sleep Disorders
- Etc.
5Subsystems
- Electrode
- Amplifier
- Multiplexer
- Transmitter
- Receiver
6 Block Diagram of System
MUX
7Possible Improvements
- Areas to improve in the neural recording system
- Electrodes
- Amplifier
- Multiplexer
- Transmitter
- Receiver
8Neural LNA Amplifier
- Functions
- Power dissipation must kept low
- Must reject noise ( flicker noise)
- Amplifying a weak signal received from the
electrodes - Must reject the DC offset created at the
electrodes-Tissue interface
9Neural Amplifier Requirements
- Power dissipation
- Bandwidth requirement
- Gain
- Stability
- Overall size
10Amplifier Solutions Design I
OTA Design
Neural Amplifier Circuit
11Amplifier Solutions Design II
Neural Amplifier Circuit
OTA Design
12Design I Simulation Results
- 10 pW power consumption
- 39.974dB gain
- Low-cutoff frequency 0.4Hz
- High-cutoff frequency 7.852kHz
- Area 0.16mm2
- Stable
13Design II Simulation Results
- 11.5 pW power consumption
- 40.362dB gain
- Low-cutoff frequency 83.114Hz
- High-cutoff frequency 32.526kHz
- Area 0.091mm2
- Stable
14Comparison of Designs
Table 1. Summary of simulation results for the
neural amplifiers
Parameter Expected Circuit 1 Simulation Circuit 2 Simulation
Supply voltage /-2.5V /-2.5V /-1.5V
Power dissipated lt5mW 10pW 11.5pW
Gain 40dB 39.974dB 40.362dB
Bandwidth 7.0kHz 7.852kHz 32.443kHz
Low-frequency cutoff 0.025Hz 0.4Hz 83.144Hz
High-frequency cutoff 7.0kHz 7.852kHz 32.526kHz
15Conclusions
- Both designs meet the requirements of
- Power consumption
- Stability
- Size
- DC offset rejection
- Design I is significantly better, due to the more
desirable bandwidth requirements.
16Implementation
17References
18Exit(0)