1558-1748 (c) 2019 IEEE. Personal use is permitted, but republication/redistribution requires IEEE permission. See http://www.ieee.org/publications_standards/publications/rights/index.html for more information. This article has been accepted for publication in a future issue of this journal, but has not been fully edited. Content may change prior to final publication. Citation information: DOI 10.1109/JSEN.2019.2933917, IEEE Sensors Journal IEEE SENSORS JOURNAL, VOL. 18, NO. 19, AUGUST 2015 1 An Open-Source Hardware Acquisition Platform for Physiological Measurements Juan A. Castro-Garc´ ıa, Student Member, IEEE, Alberto J. Molina-Cantero, Manuel Merino-Monge, and Isabel M. G´ omez-Gonz´ alez, Senior Member, IEEE Abstract—In this article we present a new low-cost physiolog- ical signal acquisition platform in compliance with Open-Source Hardware philosophy. The device contains an electrocardiogra- phy circuit; two electrodermal activity channels that implement a time-multiplexing technique to remove inter-channel coupling interference; and a skin temperature module. These features make the platform suitable for experiments involving Affective Computing. We have assessed the performance of this device in terms of Signal-to-Noise Ratio (SNR) and Background Noise Level (BNL), an adapted MPEG7 feature for electrocardiography signals. This platform outperforms other commercial devices and most similar state-of-art studies on SNR. Results show values of 11.96, 30.01, 68.40 dB for ECG, EDA and ST respectively for SNR, and -28.31 dB for BNL in ECG. Both features were obtained by sampling physiological signals at a frequency of 256 Hz. Index Terms—Affective Computing, Biomedical measurement, ECG, EDA, Skin Temperature. I. I NTRODUCTION T HE acquisition and processing of bioelectrical signals has been widely used for different purposes and research areas, such as: monitoring subjects’ health state with wearables devices [1]–[3]; disease diagnosis [4]; rehabilitation processes [5]; improving the quality of life of people with disabilities and their autonomy by facilitating the control, for example, of a wheelchair [6]; in accessing a computer through a Human Machine Interface (HMI) based on electroencephalography (EEG) [7] or other kinds of bioelectrical signal [8]; etc. These signals can also be used to detect a subject’s affective state and therefore identify emotions [9]–[11]. Two of the most common signals used in this research area are electrocardiog- raphy (ECG) and electrodermal activity (EDA) [12], although others, such as temperature or EEG, are also often included [13]. Several features, such as heart rate (HR), heart rate variability (HRV) [14] and breathing rate (BR) [15], obtained in ECG, can be related with different emotional states. For example, Dishman et al. [16] observed an inverse relationship between stress and the normalized high frequency (0.15-0.5 Hz) of HRV. Boucsein [12] observed so-called emotional sweating, produced by an increase in the sympathetic activity exerted over the sweat glands in the skin, mainly on palmar and plantar sites, due to psychological stimuli and which influences EDA signals. Skin temperature (ST) is also an J. A. Castro-Garc´ ıa, A. J. Molina-Cantero, M. Merino-Monge and I. M. G´ omez-Gonz´ alez were with the Department of Electronic Technology, Seville University, Seville, 41012 Spain. E-mail: {jacastro, almolina}@us.es, manmermon@dte.us.es, igomez@us.es Manuscript received September 26, 2018; revised August 26, 2015. indicator of sympathetic activity [17] as consequence of a response to a stimulus, such as stress or pain, which induces vasoconstriction and, thus, a decrease in skin temperature [18]. Most bioelectric signal acquisition platforms are proprietary, expensive and difficult to customize. To address this situation, several Open-Source Hardware (OSHW) [19] initiatives have emerged such as OpenEEG [20], Gamma Cardio CG [21], e-Health [22] or BITalino [23]. OpenEEG is an EEG project which contains several platform designs with up to 6 channels, a configurable sampling frequency (F s ) ranging from 100 up to 256 Hz, and RS-232 or USB connection ports. Gamma Cardio CG [21], is a 12-lead ECG with an F s of 1000 Hz and USB connection. e-Health v2 [22] is a shield for the popular Arduino platform that includes nine sensors, such as, ECG, EDA, ST, airflowmeter, glucometer, etc. However, this device is expensive (about 240 e for the shield alone) and has evolved to a non OSHW called MySignals [24]. BITalino is a comparatively low-cost OHSW (> 150 e) hardware platform for signal acquisition, based on the Arduino Uno processor. It includes four biological sensors: accelerometer, light sensor, electromyography (EMG), ECG, EDA and EEG. In this work, we present a low-cost OSHW acquisition platform called Physiological Signal Module (PSM), which is capable of acquiring useful signals in experiments on affectiv- ity. Section II describes the platform requirements and shows their schematics; it then goes on to compare PSM signals with other popular commercial devices. Section III contains the experimental and comparison methods and Section IV outlines the results. Finally, Sections V and VI present a general discussion of the results, conclusions and future work. II. THE ACQUISITION PLATFORM Affective research often employs an ECG channel and an EDA channel. However, the platform of this paper contains two EDA channels to support multiple arousal and/or asym- metric arousal theories [25]–[27], which posit the existence of left-right EDA differences due to asymmetries in brain activity in complex emotions. In the same way, we included a temperature circuit to measure ST changes in response to emotional stimuli. As a previous step in the development of the platform, we reviewed the scientific literature to find out the typical values for sampling frequency (F s ) and bandwidth (BW). Table I summarizes those values for the ECG circuit. Most researchers use commercial circuits or databases with ECG signals acquired at a typical sampling rate of 250Hz, or higher