Introduction

This R Markdown file demonstrates healthcare signal processing and monitoring using sample ECG (Electrocardiogram) data. It covers: - Generating synthetic ECG data - Signal filtering - Feature extraction (heart rate) - Visualization - Simple monitoring/alert logic

Generate Sample ECG Data

set.seed(123)
time <- seq(0, 10, by = 0.002)
# Simulate ECG-like waveform
ecg <- 1.5 * sin(2 * pi * 1.2 * time) + 0.5 * sin(2 * pi * 3 * time) + rnorm(length(time), 0, 0.2)
ecgsig <- ecg + 0.2 * sin(2 * pi * 50 * time) # Add 50Hz noise

data <- tibble(time = time, ecg = ecgsig)
head(data)

## # A tibble: 6 × 2
##    time    ecg
##   <dbl>  <dbl>
## 1 0     -0.112
## 2 0.002  0.113
## 3 0.004  0.585
## 4 0.006  0.329
## 5 0.008  0.309
## 6 0.01   0.550

Visualize Raw ECG Signal

ggplot(data, aes(x = time, y = ecg)) +
  geom_line(color = 'steelblue') +
  labs(title = 'Raw ECG Signal', x = 'Time (s)', y = 'Amplitude') +
  theme_minimal()

Filter the Signal (Remove 50Hz Noise)

fs <- 500 # Sampling frequency (Hz)
nyq <- fs / 2
bf <- butter(4, c(0.5, 40) / nyq, type = "pass")
data$ecg_filt <- filtfilt(bf, data$ecg)

Visualize Filtered ECG Signal

ggplot(data, aes(x = time, y = ecg_filt)) +
  geom_line(color = 'darkgreen') +
  labs(title = 'Filtered ECG Signal', x = 'Time (s)', y = 'Amplitude') +
  theme_minimal()

Feature Extraction: Simple Heart Rate Estimation

# Detect R-peaks (simple thresholding for demo)
peaks <- which(diff(sign(diff(data$ecg_filt))) == -2 & data$ecg_filt[-c(1, length(data$ecg_filt))] > 1)
peak_times <- data$time[peaks]
rr_intervals <- diff(peak_times)
heart_rate <- 60 / mean(rr_intervals)

cat("Estimated Heart Rate:", round(heart_rate), "bpm\n")

## Estimated Heart Rate: 448 bpm

Monitoring/Alert Logic

if (heart_rate < 50) {
  alert <- "Bradycardia detected (Low heart rate)!"
} else if (heart_rate > 100) {
  alert <- "Tachycardia detected (High heart rate)!"
} else {
  alert <- "Heart rate normal."
}
cat(alert)

## Tachycardia detected (High heart rate)!

EEG vs ECG Signal Comparison

Electroencephalogram (EEG) and Electrocardiogram (ECG) are both biomedical signals but represent different physiological processes:

ECG: Measures the electrical activity of the heart. Typical frequency range: 0.5–40 Hz. Used for cardiac monitoring.
EEG: Measures the electrical activity of the brain. Typical frequency range: 0.5–100 Hz. Used for neurological monitoring.

Below, we generate synthetic EEG and ECG signals, visualize them, and compare their characteristics using a dynamic, interactive plot with the dygraphs package.

# Generate synthetic EEG signal (higher frequency, lower amplitude)
time <- seq(0, 10, by = 0.002)
eeg <- 0.1 * sin(2 * pi * 10 * time) + 0.05 * sin(2 * pi * 20 * time) + rnorm(length(time), 0, 0.02)

# ECG signal already generated as data$ecg_filt
# Combine for comparison
df_compare <- data.frame(
  time = time,
  ECG = data$ecg_filt,
  EEG = eeg
)

# Convert time to POSIXct for xts compatibility
start_time <- as.POSIXct("2024-01-01 00:00:00")
df_compare$datetime <- start_time + df_compare$time

df_xts <- xts(df_compare[, c("ECG", "EEG")], order.by = df_compare$datetime)

dygraph(df_xts, main = "Comparison of ECG and EEG Signals") %>%
  dySeries("ECG", color = "darkgreen") %>%
  dySeries("EEG", color = "purple") %>%
  dyRangeSelector() %>%
  dyOptions(stackedGraph = FALSE, fillGraph = FALSE)

Analysis

ECG: Shows periodic, larger amplitude waves (heartbeats), lower frequency.
EEG: Shows smaller, more irregular, higher-frequency oscillations (brain activity).
Clinical Use: ECG is used for heart health monitoring; EEG is used for brain activity analysis (e.g., epilepsy, sleep studies).

The interactive plot above allows you to zoom and pan to explore the differences in signal morphology and frequency content.

Conclusion

This example demonstrates basic healthcare signal processing and monitoring in R. For real-world applications, use validated algorithms and real patient data.

Satya’s Healthcare Signal Processing and Monitoring Analytics

Satya Panda

2024-10-01