2.1.1. Population

One hundred and three subjects participated in the present study (N = 103). The group consisted of 43 males and 55 females (mean age: 61.8 years; minimum 20 years; maximum 89 years). 15 subjects had primary school education, 36 subjects had secondary school education, and 49 had higher education. Data on education level was missing for 3 subjects. The origin of the hearing loss was variable: acquired hearing loss (*n* = 83), congenital hearing loss (*n* = 20). The etiologies are described in Table 1. The category "other" includes hearing loss from otitis associated with another context, head trauma, ototoxicity, toxic shock associated with progressive deafness, drug overdose, drug treatment, sepsis, as well as idiopathic deafness, and fragile X syndrome.

Subjects were recruited from a panel of patients with severe to profound hearing loss at the Institut Universitaire de la Face et du Cou de Nice during a pre-cochlear implant assessment. All subjects included had bilateral severe to profound hearing loss and were native French speakers. This study was approved by the Recherches Non Interventionnelles de l'Université Côte d'Azur (CERNI) AVIS number 2020-62. All participants signed an informed consent form before the start of the study.


**Table 1.** Etiologies of hearing loss in the population ranked by frequency.

#### 2.1.2. Materials and Procedure

We used Cardebat's fluencies [23] to assess phonemic and semantic lexical access. The phonemic fluencies [P] and [R] and the semantic fluencies "animals" and "fruit" were used. Each subject took a phonemic fluency test and a semantic fluency test in a randomized fashion. For each fluency test, the participant had to produce as many words as possible within 2 min. The verbal fluency tests were given to the patients during the pre-cochlear implant assessment session.

#### 2.1.3. Statistical Analyses

The Z-score was calculated using Cardebat's fluency calibration [23]. This calibration takes into account the type of fluency as well as the gender, age and education level of the participants. The Z-score is the standard deviation. The closer the Z-score is to 0, the closer it is to the norm. A negative Z-score shows below average performance, and a positive Z-score shows above average performance. The Shapiro-Wilk test was used to check the normal distribution of the data. Since the data is normally distributed, a paired-sample t-test was used to compare the Z-scores of phonemic and semantic fluency. In order to compare the phonemic and semantic fluency Z-scores between acquired and congenital deafness, we used an independent samples t-test and a Mann Whitney test according to the distribution of the data. The Pearson correlation was used to establish the correlation between the phonemic and semantic fluency Z-scores. Significant results are reported as *p* < 0.05 (*p* < 0.05 \*, *p* < 0.01 \*\*, *p* < 0.001 \*\*\*).

#### *2.2. Results*

Of the 103 participants, 2 individuals were below the pathology threshold (Z = −2) in phonemic fluency and 6 individuals in semantic fluency. Table 2 shows the results of the Z-scores for phonemic and semantic fluency.


**Table 2.** Phonemic and semantic fluency results by Z-score.

The mean Z-score for phonemic fluency is higher than that for semantic fluency. The *p*-value shows that this result is significant (*p* = 0.043 \*). In addition, there is a significant positive correlation between the phonemic fluency Z-score and the semantic fluency Z-score (*r* = 0.630, *p* < 0.001 \*\*\*). Figure 1 shows the distribution of phonemic and semantic fluency Z-scores.

**Figure 1.** Distribution of Z-scores in the two fluency conditions.

The results of the phonemic and semantic fluency Z-scores according to the origin of the participants' deafness (acquired deafness or congenital deafness) are described in Table 3.

**Table 3.** Phonemic and semantic fluency Z-scores according to the origin of the deafness.


The results show higher Z-scores in phonemic and semantic fluency for participants with congenital deafness. The *p*-value is significant for the Z-score in semantic fluency (*p* = 0.043 \*). The *p*-value is not significant for the phonemic fluency Z-score.
