Skip to main content

Table 1 Comparison of different methods for motor decoding

From: Human motor decoding from neural signals: a review

 

Cortical

Peripheral

 

EEG

ECoG

Intra-cortical

Peripheral nerves

EMG

Decoding site

Scalp

On the surface of the brain

Penetrated into cortical tissues (e.g. PPC, M1)

Peripheral nerves (e.g. ulna, median, radial nerves)

Muscles

Types of electrode

Disk electrodes

Flexible electrode array

Utah array

Cuff, intra-neural electrodes

Surface electrodes,needle electrodes

Typical spatial resolution [14,169–173]

5-9 cm

<5 mm

3-5 μm

0.5-2 mm

>10 mm

Frequency spectrum

0.5-100Hz

0-500Hz

100Hz-20kHz

0.1-10kHz

0.1Hz-10kHz

Decodable intention

Movement of different body parts, 2D and 3D direction of movement, different movements of the same limb, individual finger movement

Movement of different body parts, different hand gestures, 2D position and velocity of movement, continous finger position

2D direction of movement, different hand gestures

Different hand gestures

Different hand gestures, proportional control of grasps

Signal-to-noise ratio

Low

Medium

High

Low

High

Signal feature

Bandpower, ERS/ERD

Bandpower, LMP

Spike firing rate, LFP

Action potential firing rate

Various signal features (e.g. RMS, variance, mean absolute value etc.)

Invasiveness

Low

High

Very high

Medium

Low

Advantages

Non-invasive, easily deployable

Fine-grained and robust feature, mature surgical procedures as part of epilepsy treatment

Fine-grained and robust feature

Less invasive, potentially contains detailed information about muscle activations

Non-invasive, mature technology, easily deployable

Disadvantages

Low signal-to-noise ratio, high variability of features between sessions, time-consuming to setup

Invasive, long-term implantation not common

Very invasive, require implantation surgery

Low signal-to-noise ratio

Limited DoF, exessive cross-talk between different channels