Brain-computer Interfaces

• A. Astrocytes
• B. Neurons ✓
• C. Microglia
• D. Oligodendrocytes

Why: Neurons are specialized cells that generate and transmit electrical signals (action potentials) across their membranes. Glial cells such as astrocytes, microglia, and oligodendrocytes support neurons but do not primarily carry signals.

• A. -70 mV ✓
• B. +40 mV
• C. 0 mV
• D. -150 mV

Why: A typical neuron's resting membrane potential is about -70 mV, meaning the inside of the cell is negative relative to the outside. This is maintained mainly by the sodium-potassium pump and the selective permeability of the membrane to ions.

• A. Potassium (K+)
• B. Chloride (Cl-)
• C. Sodium (Na+) ✓
• D. Calcium (Ca2+)

Why: During the rising phase of an action potential, voltage-gated sodium channels open and Na+ rushes into the cell, rapidly depolarizing the membrane. Potassium ions flow out later to repolarize the cell.

• A. Electromyogram (EMG)
• B. Electrocorticogram (ECoG)
• C. Electroencephalogram (EEG) ✓
• D. Magnetoencephalogram (MEG)

Why: EEG is recorded non-invasively from electrodes on the scalp and reflects the synchronized postsynaptic activity of large populations of cortical neurons. EMG measures muscle activity, ECoG is recorded from the cortical surface, and MEG measures magnetic fields.

• A. They are recorded only from a single neuron
• B. The skull and scalp blur and spread electrical activity before it reaches the electrodes ✓
• C. They can only detect activity below 1 Hz
• D. Electrodes must be implanted deep in the brain to record them

Why: The skull, cerebrospinal fluid, and scalp have different conductivities that smear the electrical fields generated by cortical sources, making it difficult to pinpoint exactly where in the brain the activity originated.

• A. Microelectrode arrays implanted in cortical tissue
• B. Silver/silver chloride (Ag/AgCl) electrodes placed on the scalp ✓
• C. Depth electrodes inserted into subcortical structures
• D. Carbon nanotube fibers injected into the bloodstream

Why: Ag/AgCl scalp electrodes are the standard non-invasive choice for EEG because they provide stable, low-noise electrode-skin interfaces. Implanted and injected options are invasive approaches, not standard for routine non-invasive EEG.

• A. To permanently bond the electrode to the skull
• B. To amplify the brain signal before it reaches the amplifier
• C. To reduce skin-electrode impedance and improve signal quality ✓
• D. To block electrical interference from the environment entirely

Why: Conductive gel or saline lowers the impedance between the scalp and the electrode by penetrating the high-resistance outer skin layer, allowing cleaner signal transmission. It does not amplify signals or fully block all interference.

• A. Impedance matching
• B. Analog-to-digital conversion (ADC) ✓
• C. Electrode referencing
• D. Signal grounding

Why: Analog-to-digital conversion samples the continuous voltage signal at discrete time points and encodes each value numerically, enabling digital storage and processing. The other options are preparatory steps, not the conversion itself.

• A. It measures the voltage difference between a recording site and a stable reference point, since EEG measures potential differences ✓
• B. It stores the recorded data for later analysis
• C. It increases the sampling rate of the amplifier
• D. It converts electrical signals into light pulses

Why: EEG measures voltage differences, so every recording requires a comparison point. A reference electrode, placed at a relatively inactive location, provides this baseline against which active electrode voltages are measured.

• A. They have very high amplitude, often exceeding several volts
• B. They are extremely small in amplitude, typically in the microvolt range ✓
• C. They travel faster than the speed of electrical noise
• D. They only exist during sleep

Why: Scalp EEG signals are typically only tens of microvolts in amplitude, making them highly susceptible to noise from muscle activity, power lines, and electrode drift. Their small size demands careful amplification and shielding.

• A. Single-neuron action potentials
• B. Electroencephalography (EEG) signals ✓
• C. Local field potentials recorded from deep electrodes
• D. Blood oxygen level changes measured by fMRI

Why: EEG measures electrical activity through electrodes placed on the scalp, making it non-invasive and the most widely used signal source in consumer and research BCIs. Single-neuron recordings and deep local field potentials require implanted electrodes, and fMRI measures blood oxygenation rather than electrical activity.

• A. To increase the amplitude of all frequency components equally
• B. To remove slow drifts and high-frequency noise while keeping the frequency range of interest ✓
• C. To convert the signal from the time domain to the frequency domain
• D. To synchronize signals across multiple electrodes

Why: A bandpass filter passes frequencies within a chosen range, such as 0.5 to 40 Hz for many EEG applications, while attenuating frequencies outside it. This removes DC drift, muscle artifacts, and line noise, preserving the neural rhythms relevant to the task.

• A. Delta band (0.5 to 4 Hz)
• B. Theta band (4 to 8 Hz)
• C. Mu and beta bands (roughly 8 to 30 Hz) ✓
• D. Gamma band (above 40 Hz)

Why: Mu (8 to 13 Hz) and beta (13 to 30 Hz) rhythms over the sensorimotor cortex show event-related desynchronization during imagined or actual movement. This makes them key features for motor imagery BCIs.

• A. The exact location of the neuron that generated the signal
• B. The power distribution of the signal across different frequencies ✓
• C. The subject's intention expressed in natural language
• D. The electrode impedance at each channel

Why: The FFT decomposes a time-domain signal into its constituent frequency components, revealing how much power exists in each frequency band. This spectral information is a common basis for feature extraction in BCIs.

• A. To reduce the file size of recorded data
• B. To avoid aliasing, which occurs when high-frequency components are misrepresented as lower frequencies ✓
• C. To ensure the electrodes remain in contact with the scalp
• D. To increase the signal-to-noise ratio of the amplifier

Why: This is the Nyquist criterion. If the sampling rate is lower than twice the highest frequency present, that frequency folds back and appears as a false lower frequency, a distortion called aliasing. Anti-aliasing filters are typically applied before sampling to prevent this.

• A. To stimulate neurons with electrical pulses
• B. To map recorded neural activity to an intended output such as a movement or command ✓
• C. To amplify the signal-to-noise ratio of the electrodes
• D. To store raw neural data for later archival

Why: A neural decoder learns a mapping from features of recorded brain activity to the user's intent or the variable of interest, such as cursor velocity or a selected letter. Stimulation, amplification, and storage are related but distinct functions.

• A. The average firing rate within a fixed time window ✓
• B. The total battery voltage of the implant
• C. The sampling frequency of the amplifier
• D. The impedance of the electrode wire

Why: Firing rates computed over time bins are a standard and robust feature for decoding intent from spike trains. Battery voltage, sampling frequency, and electrode impedance describe hardware properties rather than neural information.

• A. Linear models always achieve perfect accuracy on neural data
• B. They are simple, interpretable, computationally cheap, and provide a reference point for more complex models ✓
• C. They require no training data at all
• D. They can only be applied to EEG and never to intracortical recordings

Why: Linear decoders are easy to train and interpret and run fast enough for real-time use, so they serve as a benchmark. More complex models such as neural networks must demonstrate improvement over this baseline to justify their added complexity.

• A. The decoder will run too slowly for real-time use
• B. The decoder will perform well on training data but poorly on new data from the same user or later sessions ✓
• C. The decoder will automatically disconnect from the electrodes
• D. The decoder will require more electrodes to function

Why: Overfitting means the model memorizes noise and idiosyncrasies of the training set, so its performance drops on unseen data. This is especially problematic in BCIs because neural signals drift over time and across sessions.

• A. It increases the number of electrodes used by the decoder
• B. It provides a more reliable estimate of how the decoder generalizes by testing on held-out data splits ✓
• C. It eliminates the need for any test dataset
• D. It guarantees the decoder will never make errors

Why: Cross-validation repeatedly splits data into training and validation folds, giving a more robust estimate of generalization performance than a single split. It helps detect overfitting and reduces the chance that results depend on one lucky partition.

• A. Actually moving a limb while the BCI records the movement
• B. Mentally rehearsing a movement without physically executing it ✓
• C. Viewing images of machines to train visual attention
• D. Passively receiving electrical stimulation to the motor cortex

Why: Motor imagery is the mental simulation of a movement without any overt physical execution. BCIs detect the brain activity associated with this imagined movement and translate it into control commands.

• A. Functional MRI blood-oxygen signals only
• B. Electroencephalography (EEG) signals over the sensorimotor cortex ✓
• C. Electromyography (EMG) signals from forearm muscles
• D. Heart rate variability

Why: EEG records electrical activity from the scalp and captures sensorimotor rhythms, such as mu and beta rhythms, which change when a person imagines movement. EMG measures muscle activity, which is absent during pure imagery.

• A. Event-related desynchronization (ERD) ✓
• B. Long-term potentiation
• C. Cortical spreading depression
• D. P300 amplitude enhancement

Why: Event-related desynchronization is a reduction in rhythmic oscillatory power in a specific frequency band linked to cortical activation. It occurs during actual and imagined movement and is a key feature used in motor imagery BCIs.

• A. Controlling a robotic arm or wheelchair for people with paralysis ✓
• B. Improving memory consolidation during sleep
• C. Measuring blood glucose levels non-invasively
• D. Detecting emotional facial expressions in video

Why: Motor imagery BCIs can allow people with severe motor impairments, such as spinal cord injury or ALS, to control external devices like robotic arms, wheelchairs, or computer cursors using only their imagined movements.

• A. To physically strengthen the user's muscles
• B. To train the classifier on the user's unique brain signal patterns ✓
• C. To remove the user's need for attention
• D. To permanently alter the user's brain anatomy

Why: Brain signal patterns vary substantially between individuals and even across sessions. A calibration phase collects labeled examples of imagined movements so the classifier can learn that specific user's patterns and translate them into reliable commands.

• A. A headset that reads brain activity through the scalp using electrodes
• B. An interface that places electrodes directly into or on the surface of the brain, typically through surgery ✓
• C. A system that uses functional MRI scanners to decode thoughts from outside the skull
• D. A software program that predicts user intentions from keyboard and mouse input

Why: Invasive BCIs require surgical implantation of electrodes inside the skull, either penetrating brain tissue or resting on the cortical surface, which allows them to record neural signals with much higher fidelity than external methods.

• A. They provide single-neuron resolution in deep brain structures
• B. They require no surgery, making them safer and more accessible for everyday use ✓
• C. They eliminate all signal noise caused by muscle movement and environmental interference
• D. They can be permanently implanted without any risk of tissue reaction

Why: Because non-invasive BCIs sit on the scalp rather than inside the body, they avoid surgical risks, infection, and long-term tissue response, which makes them far easier to adopt outside clinical settings.

• A. It cannot detect any electrical activity from the brain
• B. The skull and scalp blur and attenuate the signals, reducing spatial resolution ✓
• C. It requires a permanent implant to record accurately
• D. It can only be used on animals and not on humans

Why: Signals generated by neurons must pass through the skull, cerebrospinal fluid, and scalp, which spreads and weakens them. This limits the spatial precision of EEG compared with signals recorded directly from the cortex.

• A. Electroencephalography (EEG)
• B. Electrocorticography (ECoG) ✓
• C. Functional near-infrared spectroscopy (fNIRS)
• D. Magnetoencephalography (MEG)

Why: Electrocorticography (ECoG) uses electrode arrays placed on the cortical surface under the skull. It is considered partially invasive: it offers better signal quality than EEG while causing less tissue damage than penetrating electrodes.

• A. Increased battery drain that shortens the device's useful life to a few hours
• B. Surgical complications such as infection, bleeding, and inflammatory tissue response around the implant ✓
• C. Inability to interpret any motor commands from the user
• D. Excessive heat generated by the scalp electrodes during normal operation

Why: Implanting electrodes requires neurosurgery, which carries risks of infection, hemorrhage, and scar tissue formation. Over time, the body's immune response can degrade signal quality, so these risks are central to evaluating invasive BCIs.

• A. The speed at which the device processes signals
• B. The privacy and sensitivity of the personal information contained in brain signals
• C. The color and shape of the electrodes used
• D. The cost of manufacturing the hardware

• A. Users agree to the device automatically after purchase
• B. Participants understand the risks, benefits, and data use before voluntarily agreeing to take part
• C. Researchers consent to publish results in journals
• D. Hospitals approve the device for clinical use only

• A. Battery overheating from wireless charging only
• B. Infection, tissue damage, or inflammation from the surgical implant
• C. Loss of internet connectivity
• D. Screen flickering during use

• A. Because brain signals can potentially reveal thoughts, emotions, or intentions that users may not wish to share
• B. Because BCIs always store data in public cloud servers
• C. Because electrodes can be seen through clothing
• D. Because BCIs require users to share passwords

• A. Developing BCIs only for gaming entertainment
• B. Creating non-invasive headsets for office productivity only
• C. Assistive and restorative BCIs for motor and communication rehabilitation
• D. Replacing all surgical procedures with mobile apps