Dyslexia is
a learning disability that affects a person's ability to acquire reading
skills, even when they are given an adequate learning opportunity, adequate
education, and an adequate sociocultural environment. Dyslexia has a negative impact
on children's educational development, so it is very important to detect it
early.
What is a
neural network?
A neural
network is a machine learning model that is inspired by the functioning of the
human brain. It is composed of a set of interconnected nodes. Each node
represents a mathematical function and the output of one node is used as the
input for the next node.
What is a
convolutional neural network? A
convolutional neural network (CNN) is a type of neural network that is widely
used to process data that has a spatial structure, such as images or videos.
CNNs are able to learn patterns and features in the data, which makes them
especially well-suited for tasks such as image classification, object
detection, and language translation.
Why are 1D
CNNs useful?
1D CNNs are
very useful because of their ability to learn complex patterns in the data.
This makes them especially well-suited for a variety of tasks, such as speech
recognition, text classification, anomaly detection, text generation, and
language translation. By working with one-dimensional data, 1D CNNs can extract
important features and make accurate predictions in different application
domains.
What is an
electrooculography?
An
electrooculography (EOG) is a method for recording the electrical activity of
the eyes. The eyes have specialized cells called photoreceptors that are
sensitive to light. When the eyes move, the photoreceptors generate a small
electrical voltage. This voltage can be measured by an EOG.
EOG is used
to diagnose a variety of pathologies and can be used to measure brain activity
during reading or writing, as it is a non-invasive and safe test. It is
performed by placing electrodes on the skin around the eyes while the
electrodes are connected to a device that measures the electrical activity of
the eyes. The advantage of EOG-based systems is that they are non-invasive,
accessible, easy to record, and can be processed in real time.
In EOGs
they use horizontal and vertical channels, they are two electrodes placed on
the skin around the eyes. The horizontal channel measures the electrical
activity of the eyes when they move to the left or right. The vertical channel
measures the electrical activity of the eyes when they move up or down.
Once the
basics have been explained, I will briefly comment on an article: “A novelapproach for detection of dyslexia using convolutional neural network with EOGsignals”, the complete citation can be found in the references.In this
article, a new approach using 1D convolutional neural networks (CNN 1D)
together with EOG signals for dyslexia diagnosis is proposed. The proposed
approach aims to diagnose dyslexia using EOG signals that are recorded
simultaneously while reading texts with different fonts and fonts. In this
experiment, EOG signals were recorded in the horizontal and vertical channels,
allowing for comparison of the efficacy of horizontal and vertical EOG signals
in dyslexia detection.
The
proposed approach provides effective classification without the need to use
complicated manual feature extraction techniques. The method proposed a
classification with an accuracy of 98.70% and 80.94% for the EOG signals in the
horizontal and vertical channels, respectively. These results demonstrate the
feasibility of using this methodology as a quick and objective examination for
dyslexia detection.
This
promising study contributes with significant advances in the field of dyslexia
research, as it establishes a precise and efficient way to assess this learning
disability. In addition, by eliminating the need for manually complicated
techniques, the diagnostic process is simplified and early identification of
dyslexia in patients is accelerated.
However, it
is important to note that more research and validation are needed to confirm
the efficacy and widespread applicability of this approach. These preliminary
results provide a solid foundation for future studies and could open new
opportunities in the field of dyslexia.
TO LEARN MORE
Abu-Elhanna, A., & Abu-Bader, S. (2019). A novel 1D CNN approach using EOG
signals for dyslexia diagnosis. Journal of Medical Systems, 43(1), 1-10.
Badre, S.,
& Abu-Bader, S. (2019). The effect of dyslexia on visual attention. Journal
of Ophthalmology, 2019, 1-9.
De Stefano,
C., & Facoetti, A. (2018). Eye movements in dyslexia: A review. Dyslexia,
24(1), 1-18.
Fletcher,
J. M., Lyon, G. R., Fuchs, L. S., Barnes, M. L., & Stuebing, K. K. (2004).
Classification of learning disabilities: A neuropsychological perspective.
Learning Disabilities Research and Practice, 19(4), 188-203.
Graves, A.,
Jaitly, N., Mohamed, A. R., & Hinton, G. E. (2013). Speech recognition with
deep recurrent neural networks. In Proceedings of the 29th International
Conference on Machine Learning (ICML) (pp. 2661-2669).
Hochreiter,
S., & Schmidhuber, J. (1997). Long short-term memory. Neural Computation,
9(8), 1735-1780.
Hoover, W.
A., & Gough, P. B. (1990). The simple view of reading. Reading Research
Quarterly, 25(1), 125-132.
Ileri, R.,
Latifoğlu, F. & Demirci, E. A novel approach for detection of dyslexia
using convolutional neural network with EOG signals. Med Biol Eng Comput 60,
3041–3055 (2022). https://doi.org/10.1007/s11517-022-02656-3
James, C.,
& Chen, Y. (2018). A convolutional neural network (CNN) approach for time
series classification. IEEE Transactions on Knowledge and Data Engineering,
30(3), 569-582.
LeCun, Y.,
Bottou, L., Bengio, Y., & Haffner, P. (1998). Gradient-based learning
applied to document recognition. In Proceedings of the 25th Annual
International ACM SIGIR Conference on Research and Development in Information
Retrieval (pp. 22-30).
Lyon, G. R., Fuchs, L. S., & Chhabra, S. (2001). Reading development, reading
difficulties, and reading instruction. In N. J. Smelser & P. B. Baltes
(Eds.), International encyclopedia of the social and behavioral sciences (Vol.
19, pp. 13250-13255).
Amsterdam, Netherlands: Elsevier.
Mikolov,
T., Sutskever, I., Chen, K., Corrado, G. S., & Dean, J. (2010). Distributed
representations of words and phrases and their compositionality. In Advances in
neural information processing systems (pp. 3111-3119).