Designing and Implementing a Portable Near-Infrared Imaging System for Monitoring of Human's Brain Functional Activity1

In this study, we used eight photodiodes (OPT101, manufactured by Burr-Brown, Tucson, AZ, which is a monolithic photodiode with on-chip single supply amplifier) and three multiwavelength light-emitting diodes (LEDs) (L4*730/4*805/4*850-40Q96-I, manufactured by Epitex, Kyoto, Japan) [1]. These LEDs send three different wavelengths at 730, 805, 850 nm, in one package (we used two wavelengths 730 and 850 nm). Since each LED sends two wavelengths, we will need six drivers for three LEDs. However, we can design only one current source; thus, need to use an analog switch to select one of the LEDs to be fed by the constant current source. We used two single-pole/single-throw (SPST) analog switches.The ATMEGA328P-PU was used to select different LEDs and wavelengths and also to control the main board. Figure 1 shows the arrangement of LEDs and photo detectors [2]. The source-detector separation in our system is 2.5 cm. Using three LEDs and eight photo detectors, we will have 24 measurements resulting in 12 channels or voxels (V1–V12). Figure 2 shows how LEDs are sequentially turned on and off [2]. First, the 730 nm wavelength of S1 (first LED) is turned ON and the other two LEDs are OFF. In this case, we read from four photo detectors around S1, giving the first measurement for V1–V4. Next, the 730 nm wavelength turns OFF, the 850 nm wavelength of S1 turns ON, and readings from four photo detectors D1–D4 give the second measurement for V1–V4. Then, both wavelengths (all LEDs and all wavelengths) turn OFF. After 10 ms, the same procedure is repeated for S2 and the four photo detectors around it (D3–D6) resulting in two measurements for V5–V8. Finally, doing the same procedure for S3 and D5–D8 gives us the values for V9–V12. We continue this routine to measure near-infrared spectroscopy (NIRS) data continuously. The duration of each wavelength emission is 10 ms.When one wavelength is ON, we read from four photo detectors simultaneously: D1–D4 for S1, D3–D6 for S2, and D5–D8 for S3. In order to switch between photo detectors, we used two multiplexers [1]. Each photo detector output is amplified and buffered followed by low pass filtering to remove high frequency noises. Finally, it is read by analog-to-digital convertor (ADC) of microcontroller. The low pass filter is a simple RC filter with cut-off frequency of 884.19 Hz. In order to record the NIRS measurements, the microcontroller sends the measurement values (read from ADCs) to a personal computer or laptop using a universal serial bus (USB) cable. Our imaging system (Fig. 3) uses rechargeable 3800 mAh Li-ion 12 V battery pack.To evaluate our designed system, we used two cognitive tasks to compare the functional brain activity during the tasks compared to rest state: the tasks were verbal fluency (VF) and color distinction (CD). In VF task (VFT), at first subjects were asked to close their eyes and recite the alphabets slowly to stabilize the prefrontal activity for 2 min; this was used as their rest state. Then, they were asked to open their eyes and upon seeing a letter on the laptop screen, say as many words as they could recall that started with the shown letter. The letter changed every 20 s; we used six letters of “F,” “A,” “S,” “P,” “G,” and “R” in that order. After 2 min of VFT, subjects were asked to close their eyes again and try to not think about anything and recite the alphabets slowly as the second rest state for 2 min. Each channels' signals were normalized to the range of [−1, +1]; then, averaged between the nine subjects who participated in this task.During CD task, a window of seven names of the colors was written with a different color as their names, and the subjects were instructed to name the color of the written color's name; for example, the word “blue” was written by color red and the subject was supposed to say the word “red” instead of reading the word “blue.” The task duration was 40 s, five subjects participated in this task.The results on average show that in both cognitive tasks in the first rest state, the changes in [HbO2], [HbR], and [HbT] are around baseline, but as the cognitive task began both [HbO2] and [HbT] increased and [HbR] decreased; this was expected. At the end of the task, when subjects closed their eyes and relaxed, all signals converged toward zero reaching the baseline (Fig. 4). This is consistent with results reported in Refs. [3–5]. Analysis for VF and CD tasks show that changes in concentration of deoxygenated hemoglobin have the lowest contrast-to-noise ratio (CNR) and the largest time delay among three different types of hemoglobin signals. Changes in [HbO2] and [HbT] had almost the same values for CNR but [HbO2] changes were faster in response to brain functional activity (smaller time delays). Also, for statistical analysis we ran paired t-test for data of each channel (rest versus task) and it showed the difference between hemodynamic signals of rest state and those of task state were significant (p < 9.9 × 10−11) in all 12 channels and for all three hemoglobin signals and in both cognitive tasks. Furthermore, there was no laterality in VFT and nor in CD task (p > 0.05) in none of the three hemoglobin signals, which is consistent with the results reported in Refs. [5,6].In conclusion, our continuous wave (CW)-NIRS system has passed the initial evaluation by two common cognitive tasks with a good accuracy as expected. As the results show, our imaging system is able to detect brain functional activities of the prefrontal cortex. Most commercially available NIRS systems are big and are not portable and they are expensive. As our system is relatively easy to use and it is portable and wearable, battery operated, and affordable, it may be useful in bed monitoring studies or gait and balance studies.

Fetched live from OpenAlex and de-inverted. Abstracts are not stored in this database: the inverted indexes are 8.6 GB of the frame’s 9.3 GB of text, and the host has 13 GB free.