Tests of the Influence of DAF (Delayed Auditory Feedback) on Changes in Speech Signal Parameters

Abstract

Contemporary phonetics and speech therapy continuously seek new techniques and methods that could contribute to improving verbal communication for individuals with speech disorders. One such phenomenon, Delayed Auditory Feedback (DAF), involves the speaker hearing their own voice with a specific delay relative to real-time speech. Although the research presented in this study was conducted on healthy individuals, it offers valuable insights into the mechanisms controlling speech, which may also apply to individuals with speech disorders. This article introduces a novel method and measurement setup, focusing on selected key speech signal parameters. To characterize the impact of Delayed Auditory Feedback (DAF) on fluent speakers, speech signal parameters were measured in 5 women and 5 men during spontaneous speech and reading. Parameters such as speech rate, fundamental frequency, formants, speech duration, jitter, and shimmer were analyzed both during and prior to the application of DAF. The results of this study may find practical applications in the field of telecommunications, especially in improving the efficiency and quality of human communication.

1. Introduction

Motor activities such as speech and articulation are highly complex operations that not only rely on precise neuromuscular coordination but also exemplify how natural language interactions support adaptive behavior in complex cyber-socio-technical systems [1]. Their proper functioning directly depends on the structure and functionality of the speech apparatus and the auditory system [2], which, when working in cooperation, enable accurate articulation of words and sentences. One example of speech disorders is stuttering, which may be caused by genetic mutations [3], atypical brain activity [4], or sensorimotor changes, including alterations in auditory information processing and speech rhythm [5]. Research on speech disorders, as outlined in the new WHO guidelines in ICD-11, emphasizes the need for detailed descriptions and classifications of speech disorders, considering their diverse symptoms and severity. This approach highlights the significance of speech disorders and facilitates a better understanding of these conditions, which is critical in globally improving the quality of life for individuals with communication difficulties [6].

Delayed Auditory Feedback (DAF) is a sensory modification characterized by a precisely defined delay in auditory feedback, which can range up to 500 ms [7]. Numerous studies in the literature have addressed the selection of appropriate delay intervals to elicit the DAF effect.

Table 1. Overview of selected studies on Delayed Auditory Feedback (DAF) and the delay intervals used. The 0 ms condition, typically serving as a control, was omitted for brevity. The table supports the selection of DAF intervals employed in the current study.

Authors	Title	Delayed [ms]
Robert A. Butler, F. Thomas Galloway	Factoral Analysis of the Delayed Speech Feedback Phenomenon [8]	85, 170, 255, 340, random
S.P. Neilson, B.D. Burke, A.J. Yates	Stability of Susceptibility to DAF [9]	80, 180, 280
Anne Gatest, John L. Bradshaw, Norman C. Nettleton	Effect of Different Delayed Auditory Feedback Intervals on a Music Performance Task [10]	10, 13, 18, 20, 27, 35, 48, 55, 65, 75, 85, 105
Bryan D. Burke	Susceptibility to Delayed Auditory Feedback and Dependence on Auditory or Oral Sensory Feedback [11]	50, 100, 200, 400
P.B. Vrtunski, Mark Martinez, Francois Boller	Evaluation of Delayed Auditory Feedback (DAF) Effect: Comparison Between Subjective Judgments and Objective Measures [12]	180, 360
T. Fukawa, S. Yoshida	Sex Difference in Susceptibility to Delayed Auditory Feedback in Oral Reading Tasks [13]	150, 200, 300, 400, 500
David M. Corey, Vishnu Anand Cuddapah	Delayed Auditory Feedback Effects During Reading and Conversation Tasks: Gender Differences in Fluent Adults [7]	180, 230, 280, 330, 380
Hideki Takaso, Frank Eisner, Richard J.S. Wise, Sophie K. Scott	The Effect of Delayed Auditory Feedback on Activity in the Temporal Lobe While Speaking: A Positron Emission Tomography Study [14]	50, 125, 200
Arghavan Bahadorinejad, Farshad Almasganj	Delayed Auditory Feedback for speech disorders [15]	25-75, 75-250
Heechong Chon, Shelly Jo Kraft, Jingfei Zhang, Torrey Loucks, Nicoline G. Ambrose	Individual Variability in Delayed Auditory Feedback Effects on Speech Fluency and Rate in Normally Fluent Adults [2]	250

presents a selection of relevant articles along with the specific delayed used. For clarity, a 0 ms condition is not listed, as its inclusion is typically implicit. These findings support the rationale behind the delay intervals chosen by the authors, as described in the Study Design section.

The article opted for delays of 180, 250, and 360 ms, as delays below 100 ms did not significantly affect the subjects. In a recent article from 2022, authors Ozker, Dolye, Devinsky, and Flinker examined auditory cortex activation along with the DAF effect for delays of 50, 100, and 200 ms. Activity in the auditory cortex (STG) increased with increasing delay—for a delay of 200 ms, neural activation, prolongation, and speech fluency impairment were most noticeable [16].

The phenomenon of Delayed Auditory Feedback remains the only known research tool capable of inducing stuttering symptoms in healthy individuals. At the same time, several studies suggest that the appropriate use of DAF in speech therapy effectively supports individuals with fluency disorders, enhancing communication quality [15]. A notable advantage of this phenomenon is its potential application in home settings, making therapy more accessible. Regular training sessions with DAF can lead to significant improvements in speech fluency [17].

Speech disorders, such as stuttering (ICD-10), are the most common fluency disorders, affecting approximately 0.76% of the population [18]. It is worth noting that discrepancies in data are reported in the scientific literature—some sources suggest that stuttering may affect between 1% and even 5% of the population, which allows for the assumption that it pertains to a small percentage of individuals [19]. Another disorder currently undergoing clear identification, cluttering (ICD-10 F98.6, ICD-9307.0), is observed in 5–16% of the population [20].

The therapies currently available are tailored to the needs of patients and depend, among other factors, on age and the underlying cause of the disorder. It should be emphasized that methods effective for cluttering may exacerbate symptoms in cases of stuttering, and vice versa. For this reason, the phenomenon of Delayed Auditory Feedback (DAF) should be approached with care, and an appropriately selected research group should be prepared to evaluate whether such a therapy could serve as a beneficial tool for speech therapists in supporting the primary treatment for either group.

2. Tools Used

The study used headphones with foam ear cushions to provide participants with acoustic isolation from potential ambient sounds. These headphones are closed which feature surround sound. In addition, a foam-cushioned microphone was used separately to reduce breath sounds and suppress blast consonants. The microphone used had a cardioid characteristic, sampling at 96 kHz and a bit rate of 24 bits, with a frequency response of 20 Hz to 20 kHz. The cardioid microphone was positioned 10–15 cm from the speaker’s mouth. This range was chosen so that any movement away from the microphone would be minimal and would not have a negative impact on the observations. It is worth noting that for the purposes of this study, headphones with a built-in microphone function would suffice.

Figure 1 shows a diagram of the system used in the study. The upper part of the diagram is responsible for recording audio for later analysis for the purposes of this work, while the lower part is responsible for the auditory delay effect of feedback.

Delay simulation was performed using the DAF application, which allows fine-tuning of the desired delay. This application has the ability to set the delay and record the progress (Figure 2). By its accessibility, it was decided to use this application, as anyone interested in the DAF effect can use it anywhere. Because of its simplicity of use, it is an ideal tool for working on self-improvement in both the level of correctness and intelligibility of speech. After pressing a button, the application starts its work, and everything the user says is repeated after a certain pre-selected time. Each user has the ability to record his or her speech, which is later transferred to a library of recordings, which makes it possible to analyze speech and decide what still needs to be worked on. An additional feature of the application is the ability to record sessions, which allows for later analysis of progress.

Audacity was used to record the speech of the study participants. It is an audio editing and recording program. It allows audio files to be saved in multiple formats and be recorded at multiple sampling rates. In addition, once the study was completed, the program had the ability to extract the marked sound and automatically measure its timing. For phonetic analysis, on the other hand, the Praat program was used, which was developed by researchers from the Institute of Phonetics at the University of Amsterdam—Paul Boersma and David Weenink [21]. The program is intuitive, easy to use, and at the same time allows the provision of much needed information about sound. It not only educates laymen who are taking their first steps on the subject of speech acoustics, but it is also a professional tool for researchers, which allows them to extract a lot of information on signal parameters [22,23].

3. Study Design

The study group consisted of university students (mean age: 22 years), all of whom were native speakers of Polish. The participants had no speech, hearing, vision, or neurological disorders. All were open to participating in the study and provided their consent.

As required by the Polish legislation [24], written informed consent was obtained from all participants prior to their involvement in the study, including consent for the processing of personal data. The participants were provided with comprehensive information regarding the study’s objectives and scope. The research carried no health risks, and participation was strictly voluntary [25].

The investigation examined delays of 0 ms, 180 ms, 250 ms, and 360 ms, where 0 ms represented natural speech. The participants wore headphones that isolated them from external sounds, focusing their attention entirely on auditory sensations. The tasks assigned to the participants included engaging in short conversations, describing a picture, and reading—both in natural speech and under delayed auditory feedback conditions.

Before the study began, the participants were informed about the procedure and provided with guidelines to ensure smooth execution. They were instructed to maintain a consistent distance from the microphone and avoid deviating from the specified distance between their mouths and the transducer. They were asked not to laugh during their speech, to speak loudly and clearly, maintain their natural pace, avoid rustling the paper containing the text, and refrain from tapping or stomping.

Recognizing that the study could induce stress in the participants, a stress-relief tool in the form of an elastic band was provided. This allowed the participants to channel their emotions without disrupting the study. The band did not introduce motor noise, and hand movements were not recorded in the acoustic signal. These measures were designed to prevent extraneous sounds that might interfere with the recorded speech samples. The participants were reassured that speech difficulties arising under delayed auditory feedback were expected, ensuring they did not focus excessively on verbal impairments.

The study began with spontaneous speech to familiarize the participants with the experiment’s structure. This was followed by Delayed Auditory Feedback (DAF) with delays of 180 ms, 250 ms, and 360 ms, applied in random order to prevent the participants from predicting the delay or its level of difficulty. A short conversation was conducted at the start of the test, focusing on natural topics familiar to the participant. The dialog was designed to encourage a conversation rather than interrogate the participant, helping to alleviate stress and “warm up” the speech apparatus. This section was not analyzed, as its purpose was to prepare the participants for subsequent stages of the study.

The next stage involved describing photographs selected to minimize ambiguity about their content. Participants were not informed beforehand about the images and were required to describe them spontaneously as the visual stimuli were presented in a randomized sequence. The images used in the study were frames from cartoons familiar to the participants, ensuring contextual recognition. Each of the photos described was taken into account when analyzing the results. Before this task, the participants were advised to provide descriptions as detailed as possible for a later analysis. This part aimed to assess creativity, unconventional ideas, and the ability to divide attention during spontaneous speech.

The final task involved reading two texts: a poem and a book advertisement. The poem consisted of 110 words and 209 syllables, with a Flesch–Kincaid Grade Level (Fog Index) of 6, indicating that the language was very simple and appropriate for primary school students. The book advertisement contained 139 words and 295 syllables, with a Fog Index of 9, classifying it as simple and accessible. The Fog Index is widely recognized as an international standard for assessing text complexity, including materials written in Polish [26]. Currently, there is a lack of Polish language tools that are as thoroughly tested and validated in the scientific literature. However, existing research confirms that the Fog Index can be effectively applied to Polish language texts and that the results show good correlation with other established readability assessment methods [27]. The texts were chosen to be non-demanding to ensure the participants could read them without difficulty, avoiding errors such as misreading words or substituting them with others.

The participants were instructed to first read the poem, followed by the advertisement. This sequence was intentional, as the poem’s solemn tone encourages slower, more deliberate reading, allowing researchers to examine the focus and attention during the reading tasks. The poem’s structure, which naturally slows down speech, was used to determine whether the imposed delays would further affect the reading pace. In contrast, the book advertisement was a standard text used to measure word articulation speed.

4. Result and Discussion

Articulation rate is a commonly used tool for studying speech disorders [28]. Speech tempo varies across languages. For adult Polish speakers, it averages at 2 syllables per second during slow speech and 4–5 syllables per second during faster conversations [29].

Both female and male participants demonstrated a decrease in speech tempo as the delay increased. The highest average speech tempo was observed under natural speech conditions (0 ms delay), while the lowest occurred at the maximum delay (360 ms). The women exhibited generally slower speech rates compared to the men across all delay conditions, which may suggest differences in adaptive strategies or sensitivity to DAF. The men showed less variability in speech tempo across the different delay conditions, indicating greater stability in coping with delayed auditory feedback.

Speech tempo also varied depending on the complexity of the photographs being described, which could reflect their intricacy or the individual difficulty level the participants experienced in describing them. Notably, as the delay increased, the participants’ creativity and the length of their utterances significantly declined. In the context of verbal creativity, the authors refer to the ability to produce words with a greater number of syllables, enhanced speech fluency, as well as the temporal characteristics and manner of speech production.

The results presented in Figure 3 and Figure 4 indicate that delayed auditory feedback effectively disrupts speech fluency, compelling the participants to adopt a slower and more controlled speaking style while simultaneously impacting their focus. The decrease in speech tempo at higher delays likely stems from the need for more precise control over speech to mitigate disruptions caused by delayed feedback.

The experiment focused on analyzing the fundamental frequency (pitch) of vowels. For all vowels, a general trend was observed: the frequency of the fundamental pitch changed depending on the DAF delay. In the absence of delay (0 ms), the pitch was typically the highest, while increasing the delay caused a reduction in pitch frequency, suggesting that DAF affects voice stability control.

Among the vowels analyzed, the vowel ‘A’ exhibited the highest fundamental frequency, while the vowel ‘U’ had the lowest. These results reflect articulation characteristics and reduced resonance. The vowel ‘I’ demonstrated the highest pitch stability regardless of the delay value, indicating that it is less susceptible to disruptions caused by DAF. The greatest pitch variability was observed for the vowel ‘A,’ showing significant fluctuations. The results for both groups of subjects are presented in

Table 2. Results for vowel ‘A’ in the Polish word “las” (ang. Forest) in women (w).

	Las
	Delay	Laryngeal Tone	Formants				Jitter	Shimmer	Time
	[ms]	[Hz]	1	2	3	4	[%]	[%]	[s]
w1	0	231.88	827.60	2586.28	2586.28	3888.09	0.644	3.084	0.210
	180	209.72	744.25	2603.53	2603.53	3706.12	0.584	3.604	0.192
	250	225.32	734.26	2650.14	2650.14	3793.15	0.887	2.954	0.150
	360	215.23	822.90	2747.96	2747.96	3837.26	1.180	5.319	0.154
w2	0	209.49	852.68	1521.67	2582.34	3895.52	0.453	4.987	0.289
	180	203.25	874.09	1686.67	2833.92	4554.52	0.374	5.295	0.307
	250	215.41	874.62	1685.21	2874.62	3444.65	1.326	6.781	0.351
	360	210.33	878.36	1683.33	2714.50	3967.03	0.457	7.910	0.136
w3	0	222.48	836.91	1478.21	2673.31	3120.44	0.654	10.684	0.245
	180	205.01	979.40	1579.97	2724.05	3440.65	1.169	11.458	0.291
	250	227.87	923.11	1607.13	2932.37	3482.39	1.123	14.065	0.197
	360	224.14	947.85	1629.63	2886.72	3547.17	1.421	16.163	0.238
w4	0	223.69	834.72	1500.41	2685.58	4048.66	0.473	5.831	0.172
	180	205.64	859.87	1576.31	2745.99	4326.92	0.799	5.099	0.156
	250	264.17	811.68	1590.03	2603.31	4575.02	2.101	10.053	0.186
	360	248.93	696.19	1490.29	2722.67	4560.49	1.289	11.431	0.149
w5	0	212.77	845.44	1545.45	2567.22	3221.95	1.027	8.227	0.187
	180	207.58	867.92	1572.10	2613.94	3569.74	0.695	10.230	0.182
	250	226.31	842.99	1601.20	2603.45	3477.23	1.216	11.682	0.196
	360	229.74	861.02	1677.12	2760.74	3967.53	1.120	8.596	0.192

and

Table 3. Results for vowel ’A’ in polish word “las” (ang. Forest) in men (m).

Las
	Delay	Laryngeal tone	Formants				Jitter	Shimmer	Time
	[ms]	[Hz]	1	2	3	4	[%]	[%]	[s]
m1	0	116.79	541.01	1189.92	2474.82	2955.15	0.909	4.964	0.158
	180	113.35	573.27	1179.77	2563.03	3529.23	7.413	15.470	0.140
	250	123.39	555.42	1157.77	2424.12	2974.53	6.438	10.250	0.165
	360	107.31	482.56	1160.83	2538.48	3598.16	3.880	25.263	0.242
m2	0	113.79	457.01	1360.08	2673.05	3694.22	1.200	6.120	0.073
	180	119.76	541.27	1351.85	2871.71	3562.36	1.255	8.041	0.163
	250	104.45	472.93	1545.38	2768.56	3550.25	2.419	8.090	0.169
	360	119.03	532.41	1222.55	2557.55	3503.26	0.802	4.761	0.164
m3	0	157.57	567.94	1048.12	2465.64	3795.60	4.399	18.416	0.125
	180	107.16	525.67	1238.33	2538.77	3691.86	3.421	14.957	0.259
	250	140.25	513.59	1287.81	2502.87	3945.33	2.356	14.785	0.169
	360	169.14	579.84	1311.17	2528.19	4032.85	3.373	12.748	0.205
m4	0	141.98	731.02	1239.86	2684.84	3640.49	2.136	9.960	0.104
	180	159.62	650.70	1231.50	2577.45	3252.22	0.814	3.739	0.268
	250	166.80	645.71	1233.04	2722.86	3354.40	0.974	6.647	0.176
	360	150.21	394.86	1157.29	2695.02	3486.30	0.737	5.750	0.159
m5	0	120.74	660.72	1387.29	2428.00	3901.29	0.635	7.692	0.245
	180	126.30	686.06	1398.44	2589.61	3943.21	0.573	15.494	0.183
	250	134.61	643.96	1379.29	2509.08	3983.96	0.734	12.666	0.324
	360	163.91	667.37	1437.48	2568.83	3790.44	1.247	13.538	0.147

below.

The data indicate that the group of women tested was more affected by the delay compared to the men, which contrasts with the existing literature [7]. For accurate sound production, vowels should be articulated effortlessly and without strain on the larynx. The typical fundamental frequency range for men is 85–180 Hz, and 165–255 Hz for women. The results from the existing literature and several factors may account for this discrepancy, ranging from the limited sample size to possible differences in task design, stimulus characteristics, or even cultural and linguistic background of the participants. It is also possible that individual variability within the female group was higher, leading to greater sensitivity to the delay condition. Further research with a larger and more diverse sample would be necessary to clarify the underlying causes of this divergence.

DAF-induced delays influence the pitch of the women, causing a reduction in frequency at higher delay values. At a delay of 360 ms, all vowels exhibited a decrease in pitch, which may indicate difficulties in maintaining controlled voice production under the maximum delay conditions. The 180 ms delay appeared to have the least impact on the pitch, suggesting it may represent a threshold value where the DAF effect on voice remains minimal.

Monitoring fundamental frequency is essential for accurate sound production and speech intelligibility. It also reflects emotional changes such as stress, tension, or uncertainty.

Formants are acoustic resonances generated during speech production, representing frequency bands with increased energy in the sound spectrum. These resonances are shaped by the articulatory system, including the throat, oral cavity, and nasal cavity. Special attention is given to the first and second formants: the first is related to the oral cavity, while the second reflects the tongue position. Other formants are associated with individual voice characteristics, such as timbre. Naturally, women exhibit higher formant values compared to men.

The process of formant estimation in Praat is based on spectral analysis of the speech signal and the identification of resonant frequencies corresponding to local maxima in the amplitude spectrum. A crucial part of this analysis involves both autocorrelation and Fourier transformation techniques, applied within moving time windows.

Initially, the speech signal is divided into short time frames. Each frame is then multiplied by a window function, $\omega \left(t\right)$, using Hamming window, which helps minimize edge effects and improves frequency resolution.

$$a\left(t\right)=\left(x\left(t_{mid}-{\displaystyle \frac{T}{2}}+t\right)-{\mu }_x\right) \mathrm{\ast } \omega \left(t\right)$$

(1)

$t_{mid}$—center of the analyzed frame;

$T$—window length;

$t\in \left[0,T\right]$—sliding variable within a window;

${\mu }_x$—average signal value in the analyzed fragment.

Next, the normalized autocorrelation $r_a\left(\tau \right)$ function is computed, allowing for the identification of dominant periodicities in the signal. The peak values of this function correspond to the dominant resonant frequencies.

$$r_a\left(\tau \right)=r_a\left(-\tau \right)={\displaystyle \frac{{\int }_0^{T-\tau }a\left(t\right)a\left(t+\tau \right)dt}{{\int }_0^T{a}^2\left(t\right)dt}}$$

(2)

The implementation requires performing a Fourier transform followed by increasing the resolution of the frequency domain.

$$\tilde{a}\left(\omega \right)=\int a\left(t\right)e^{-i\omega t}dt$$

(3)

The next step is to compute the inverse Fourier transform of the power spectral density, which enables a transition to the delay domain.

$$r_a\left(\tau \right)=\int {\left|\tilde{a}\left(\omega \right)\right|}^2{e}^{i\omega \tau }{\displaystyle \frac{d\omega }{2\pi }}$$

(4)

In the article by Paul Boersma, the creator of Praat, two methods of pitch analysis are described—zero-phased and cepstral—along with a discussion of their differences [30]. However, the current version of Praat employs an algorithm based on the Linear Predictive Coding (LPC) model. This method relies on the autocorrelation of the speech signal and models it as a resonant system, thereby indirectly utilizing spectral information.

Regarding formant stability, the men demonstrate greater susceptibility to changes under DAF conditions, particularly for the vowels ‘I’ and ‘U’. In contrast, women display more stable voice modulation under the same conditions, as shown in the graphs below (Figure 5 and Figure 6).

It is important to note that the obtained results are fully reliable, as no artificial conditions were imposed, requiring the participants to sustain vowels in isolation for several seconds. Instead, the analysis focused on vowel segments extracted from spontaneous speech recorded during the description of photographs. This data collection approach preserves naturalness and enhances the validity of the results within the context of the study’s conditions.

An analysis was conducted on the articulation time of 71 words repeated by each participant. The results showed that for the DAF delays of 250 ms and 360 ms, the articulation time increased for both women and men. In contrast, at the delay of 180 ms, the articulation time was shorter. Notably, the 180 ms delay not only reduced the articulation time but also increased the number of syllables in the overall speech compared to the 0 ms condition (no delay). This observation suggests that the participants were able to articulate more content under a shorter amount of time, indicating improved focus and greater speech efficiency.

The study also examined the number of syllables spoken during the description of the photographs. The results indicate that the number of syllables reflected the participants’ creativity and their ability to focus on the task or susceptibility to distraction. These data enabled the determination of speech tempo, which, for both women and men, was significantly higher at the 180 ms delay compared to the speech without a delay. These findings suggest that a 180 ms DAF delay may have a positive impact on speech fluency, making it a potentially useful tool for therapy aimed at improving speech fluency. It should be noted that the earlier conclusion is based on a sample of typical speakers. Therefore, further research is necessary to confirm these findings within clinical populations affected by speech disorders.

Additionally, jitter and shimmer parameters were analyzed. Jitter refers to irregularities in pitch, while shimmer measures the variation in acoustic wave intensity from one cycle to the next. Both parameters are determined by the characteristics of the vocal folds and are used to assess the extent to which a vowel in speech indicates voice pathology. In Praat, jitter is calculated based on the actual waveform of the voice signal rather than on a simulated model. It represents the average relative difference between successive pitch periods, thus reflecting the extent to which the duration of vocal fold vibrations varies over time.

$$Jitter={\displaystyle \frac{1}{N-1}}\sum _{i=1}^{N-1}{\displaystyle \frac{|T_i-T_{i+1}|}{\overline{T}}}$$

(5)

$T_i$—duration of the i-th period;

$\overline{T}$—average duration of periods.

Analogous to jitter, shimmer refers to variations in amplitude between consecutive vocal cycles. It reflects the degree of instability in vocal intensity and is used to assess irregularities in voice signal amplitude over time.

Jitter is considered pathological at values exceeding 1.040%, and at values above 3.810% for shimmer. However, the reliability and utility of these parameters are debated, as they are susceptible to technical factors such as microphone interference, microphone quality, software type, and background noise. These circumstances can significantly influence jitter and shimmer measurements [31].

The analysis of the impact of text complexity and Delayed Auditory Feedback (DAF) on speech tempo indicates that both factors significantly influence reading fluency. The data show that as the delay increases, the speech tempo slows down. Notably, at the 250 ms delay, participants in both groups experienced clear difficulties in articulatory fluency. This finding suggests that delays of this magnitude significantly disrupt the natural speech process, leading to reduced speech tempo and extended reading times.

As illustrated in Figure 7, the men demonstrated better performance under the DAF conditions in terms of speech tempo compared to the women, which may indicate their greater tolerance to disruptions. However, the effect of delay was equally problematic for both groups when reading texts with lower complexity levels (FOG = 6). This suggests that speakers dedicate more attention to articulation control in simpler texts.

Importantly, the observed changes in speech tempo may affect text intelligibility, which is critical for effective communication. The results indicate partial adaptation by both women and men to the conditions of disrupted auditory feedback—particularly at the 360 ms delay—when reading texts with higher complexity levels. This adaptation highlights the potential for individuals to adjust to challenging speech conditions, though the process remains influenced by the complexity of the material being read.

5. Conclusions

The effect of auditory feedback is one of the phenomena that impacts healthy individuals by introducing speech disruptions. However, the ambiguity of the currently available findings prompted the authors to explore this topic further [7]. Auditory feedback significantly affects speech fluency and production, making it a crucial tool in working with individuals who have severe speech disorders as well as those capable of fluent speech (without disfluencies).

In summarizing the collected data, it can be concluded that DAF influences healthy individuals by slowing their speech and requiring greater focus when articulating subsequent words. The authors paid special attention to the results of multiple parameters, as they highlighted the participants’ susceptibility to delays. Observations suggest that the DAF effect significantly impacts the speech tempo and reading time, which could be relevant for speech therapy aimed at improving the articulation of both healthy individuals and those with disfluencies. This paper introduces a novelty by applying DAF in healthy individuals with specific speech parameters. The proposed approach improves the understanding of speech pathologies and may contribute to more effective therapeutic interventions within the target research field.

To expand the research on Delayed Auditory Feedback in the future, it would be beneficial to include a larger group of participants. With the support of speech therapists, extending the study to include younger individuals should be considered. To develop therapy for individuals with disfluencies based on the phenomenon of DAF, it would be valuable to also examine individuals at risk of or already affected by stuttering, Parkinson’s disease, or Alzheimer’s disease [32,33]. It is worth considering expanding the research to include physical parameters such as rapid movements of the face or skin during the examination. It is also worth checking whether these movements, e.g., rapid changes in facial expression, can be correlated or used to study this type of pathology together with DAF parameters.

During the course of the study, certain limitations were identified, including the characteristics of the sample, methodological constraints, measurement precision, and the size of the study group. It would be advisable to address these issues in future research to further improve the methodological approach.

In the future, a more precise investigation of jitter and shimmer parameters using advanced software would be desirable. Additionally, to better understand the impact of the DAF phenomenon on reading, selecting materials with a balanced number of syllables and a broad range of readability indices would help assess how text complexity affects participants.