The COVID-19 disease, caused by the Severe Acute Respiratory Syndrome coronavirus 2 (SARS-CoV-2) virus, becomes one of the most fatal diseases that have ever affected human life and global economies, which has been reported in ranges from mild to severe in clinical representation [1]. The disease does not only have respiratory symptoms but also shows symptoms related to many organ systems, including the nervous system [2]. Clinical reports showed that neurological symptoms had presented a significant frequency of extrapulmonary manifestations [3]. For example, headache is given between signs in 13.6% of total COVID-19 patients with range means ranging between 4 and 24%) [4]. In addition, myalgia has shown a frequency of 20.4%, with an average ranging between 11 and 44%. [5] Furthermore, anosmia has been reported in the range of 5–88% [6]. In addition, some studies reported the frequency of neurological manifestations in severe illness cases of COVID-19 patients as 84%, while it was 36.4% in hospitalized cases [7,8]. In these studies, neurological manifestations showed up more frequently in severe cases. However, the prognoses’ influences are still not apparent [9].

In addition, surveying dozens of research works in the literature [10,11,12,13,14,15,16,17,18,19,20] showed a significant presence of neurological symptoms in COVID-19 patients, including mild and severe cases. For example, headache has been reported in 70–80% of all cases; anosmia–dysgeusia showed up in 50–60% of total cases; myalgia presented in 40–45% of all cases; fatigue was there in 30–35% of the surveyed cases; dizziness was recorded in 30–35% of patients; and 0–10% of subjects showed noncommon symptoms like numbness, migraine, loss of concentration, and seizure. The obvious correlation between some neurological symptoms and COVID-19 with variable ratio and weight can be utilized to predict the likelihood of having COVID-19 or not based on neurological symptoms. Therefore, this study analyzes thousands of COVID patients with neurological symptoms reported in the literature by surveying a couple of related research works to achieve twofold objectives. First, it determines statistically the weight of each neurological symptom in COVID-19 cases. Second, the obtained statistics of neurological manifestations are used to design the ruleset of a fuzzy logic-based model for predicting the likelihood of having COVID-19 based on neurological symptoms. To sum up, the survey of dozens of related research works yields thousands of records showing the neurological symptoms that appeared in COVID-19 cases. Then, the weight and percentage ratio of these symptoms are statistically calculated. Afterward, these data are utilized to form the ruleset of a fuzzy-logic-based system. As a result, the fuzzy logic system can decide the likelihood and certainty of having COVID-19 or not for patients based on their neurological symptoms.

In this study, the data of thousands of COVID-19 patients, including their neurological symptoms, have been collected by surveying dozens of related works in the literature. For example, in one study [10] of the surveyed literature, a group of 107 confirmed COVID-19 patients who do not require admission to the hospital were surveyed continuously for ten sequence days, reporting all their symptoms, including neurological and non-neurological symptoms, based on the predesigned questionnaire. The survey conducted for this study was performed in the city of Benisuef in Egypt. It started in the middle of June 2020 and ended in July 2020 [10].

Table 1 lists the demographic data of the patient group participating in one study [10] that has been surveyed for collecting data for this research, including the distribution of the central and peripheral neurological manifestations.

The collected data of the patients included but were not limited to age, gender, the onset of symptoms, diagnostic method, and location, i.e., hospital, clinic or institution of diagnosis, history of chronic illness, such as hypertension, diabetes, pulmonary disease, cardiac disease, history of headache, epilepsy, stroke, or dementia. The survey included ten neurological symptoms during the first 10 days of COVID-19 symptoms. That included headache (regarding its type and criteria), loss of smell and taste, myalgia, dizziness, and encephalopathy (delirium, confusion, and disturbed attention). In addition, seizures, numbness, and fatigue were registered. On the other hand, the survey included non-neurological symptoms. That was the febrile onset and pulmonary or gastrointestinal symptoms. Laboratory results were collected via e-mail, i.e., leukocyte count, C-reactive protein, D-dimer, serum ferritin, and reports of CT chest imaging. All patients consented to share their data during the period of home isolation after diagnosis of COVID-19 infection. In another study11 of the surveyed literature for collecting data for this research, Table 2 lists the demographics and comorbidities by SARS-CoV-2 result of 100 cases, including positive and negative COVID including the neurological and non-neurological symptoms manifestations. Similarly to the listed data in Tables 1 and 2, data have been collected from dozens of surveyed related research work 11–20 to be analyzed and designing the ruleset of the fuzzy logic system that will be used in forecasting the likelihood of COVID-19 based on the neurological symptoms.

The collected data from surveyed-related research work [11,12,13,14,15,16,17,18,19,20] have been analyzed using SPSS version 20.0. After the data have been analyzed statistically, the input data to the statistical analysis include the neurological and non-neurological symptoms with COVID-19 positive and negative cases, as listed in Tables 1 and 2. The outcomes of that analysis represented the frequency of neurological and non-neurological manifestations associated with COVID-19, representing the weight and percentage ratio of the symptoms. This weight and percentages will be used to design the rulesets of the fuzzification process of a developed fussy logic model, as listed in Table 3. It tabulates all input and output parameters of the fuzzy model and the related member function. The table lists the inputs and output parameters of the process in the first column. For example, the COVID-19 likelihood is an output of the fuzzification process, while all symptoms, such as headache, migraine, etc., are the inputs. The output and input are represented as percentages of certainty. The second column lists the possible values of the membership functions. For example, the production of COVID-19 likelihood could be yes if its certainty is greater than 50% but No if its likelihood is less than 50%. The third column shows the type of parameters as an input or output.

The fuzzy logic model aims to forecast the likelihood of COVID-19 in patients with such neurological symptoms, as depicted in Fig. 1. It shows the frequency of each symptom of the neurological and non-neurological symptoms that will be fed inputs to the fuzzy system on the left-hand side. In the middle, the rulesets interpret the input symptoms into the output on the right, which is the likelihood of COVID-19.

Fuzzy logic model for neurological symptoms of COVID-19

Figure 2 shows the design of the member functions of all inputs and outputs of the COVID-19 design. For example, headache is an input to the system that varies from 0 to 100%. That represents the headache severity. In addition, it shows the ranges for low, medium, and high severity of the headache the patient suffers from. Based on the inputs of these symptoms, the system can determine the likelihood of COVID-19.

COVID-19 fuzzy logic system design