Stroke is the leading cause of death and disability worldwide (Feigin et al., 2014; Mozaffarian et al., 2016), and ischemic stroke is the most prevalent type of stroke (Moskowitz et al., 2010). Moreover, the prevalence of ischemic stroke has increased among adolescents and young adults (Mozaffarian et al., 2016). Thus far, the only effective therapy for ischemic stroke is recombinant tissue plasminogen activator (Chamorro et al., 2016; Khandelwal et al., 2016). However, this therapeutic approach is effective only within a critical time window (approximately 3–4 h after stroke occurs), and is limited (Moretti et al., 2015). Therefore, the development of novel drugs and approaches that extend the therapeutic window for ischemic stroke is crucial, and an extensive study is required.

Several rodent models of ischemic stroke have been developed and adopted for research (Woodruff et al., 2011). Middle cerebral artery occlusion (MCAO) in mice or rats is a popular model for mimicking focal ischemia in the brain (Liu and McCullough, 2011; Fluri et al., 2015). Following MCAO, the classic method utilizes 2,3,5-triphenyltetrazolium chloride (TTC) as a marker of ischemic damage to the brain tissue (Bederson et al., 1986; Türeyen et al., 2004). TTC reacts with intact mitochondrial oxidative enzyme systems and is subsequently reduced by the enzymes, which leads to the normal brain tissue being stained deep red in color (Straus et al., 1948; Bederson et al., 1986). In ischemic brain tissue, damaged mitochondria with impaired oxidative systems are incapable of reducing TTC (Bederson et al., 1986; Isayama et al., 1991); thus, the corresponding damaged brain tissue remains uncolored and easily distinguishable by the naked eye or algorithms. After experimental ischemic stroke modeling and TTC staining, accurate, and rapid quantification of the location and size of the infarct is critical for intervention selection. A widely accepted method for quantification of the infarct size in TTC staining utilizes the Image-Pro Plus or Image J, which requires software licenses or substantial manual operation.

In the current study, we present the semi-automated analysis of the TTC staining (SAT) software program, an open-source, small sized (approximately 18 KB), free to distribute and use software program (for Windows 7, Windows 8, and Windows 10), which provides semi-automated processing, analysis, and visualization of data of brain sections with TTC staining. We provide the developing principle and a detailed description of the user’s guide for the software. We also aim to assess the utility of the SAT program in the analysis of infarct sizes in rodent brain sections.

To facilitate the analysis of TTC staining of brain slices after MCAO and provide alternative choice for researchers, we sought out to develop a software, which is called SAT (semi-automated analysis for TTC staining). When utilizing the SAT, the MCAO model and TTC staining were conducted as usual, and the pictures of brain slices with TTC staining were captured under the specified condition (Figure 1, 2). To easily distinguish background and the brain slices by the algorithm, the background of the brain slices under capture should be non-white as well as non-red. Hence, we recommend that the user takes advantage of black plate as a background (Figure 1, 3). Since white background was used to present TTC staining figures in a large body of studies, it is worth to note this before calculation the infarct size by the SAT program. Following the protocols described in the Figure 3, the infarct size, which is expressed as a percentage of the whole brain slice, will be appeared in the software in real time (Figure 3). However, this software is not able to export the data as TXT format or Excel format after analysis at this time.

In the present study, we performed MCAO and TTC staining to assess the accuracy and efficacy of SAT analysis of rodent brain sections by comparing the SAT results to the Image J results. We found that the SAT results were similar to the Image J results, and exhibited no significant differences (Figure 4A,B, p = 0.7721, t = 0.2976, df = 10). To further evaluate the utility of SAT analysis, we evaluated the relationship between these two methods. We observed a high level of correlation of the two data sets from the two methods (Figure 4C, p ≤ 0.0001, R = 0.958), suggesting that SAT analysis is comparable to Image J analysis.

In this study, we presented the protocol and guidelines for the use of our free software (SAT) for the analysis of brain sections with TTC staining. The development of the software was based on the brightness binarization method. Thus far, SAT has focused on the analysis of brain sections with TTC staining. In the near future, we hope it will be expanded to other applications, such as analyzing Nissl staining of brain sections after experimental stroke.

For the current version of SAT, we recognize that the data obtained by SAT were not able to export to TXT or Excel format. Moreover, analysis of TTC staining by SAT still requires manual supervision and operation. In addition, SAT can only function effectively when the image was captured on non-white and non-red backgrounds, but without the prerequisite of a specific photo capture system. SAT is capable of performing analysis with images captured by cell phones or scan systems. In addition, we used SAT to analyze several TTC-stained sections from previously published studies (see supplemental movie 1 for an example), thus suggesting the wide application of the software and indicating the quality of the previous published images as sufficient for analysis. In this study, we only applied SAT for analysis of TTC-stained mouse brain sections, but the software can also be applied to rat brain sections with TTC staining.

Recently, Macro in Image J has also been conducted in a semi-automated manner in the analysis of brain sections with TTC staining (Friedlander et al., 2017). SAT is not intended to replace Image J in the analysis of brain sections with TTC staining. However, we consider SAT to be a user-friendly and small sized software that provides an alternative analysis option for researchers in the field. Additionally, it benefits data reproducibility when the result is independently analyzed by two distinct methods.

In the future, we will aim to improve the SAT in four aspects. First, we will add the exporting function to the software, making the software capable of exporting data after analysis. Second, we are going to make the SAT able to concurrently analyze several brain sections, e.g., five sections, in a row or column. This parallel analysis may further reduce the time required and number of operational steps. Third, we are planning to develop this software as a smartphone application, making this algorithm easier to access and use. Fourth, we believe that deploying a deep learning algorithm, such as a convolutional neural network for the analysis, will improve our software in the future. By achieving the improvements mentioned above, we believe that the SAT will undoubtedly save user’s time, and result in the significant promotion of experimental studies on ischemic stroke.

X-FS, WL, and FC conceived the study. X-FS, HA, WL, and FC designed and analyzed the data. HA and WL performed the MCAO experiments and TTC staining. FC developed the software. X-FS, WL, and FC wrote the manuscript. All authors reviewed the results and approved the final version of the manuscript.

The authors declare that the research was conducted in the absence of any commercial or financial relationships that could be construed as a potential conflict of interest.