Linguistics Essay Sample Paper on A critical review of PET Experiments of Phonological Processing

A critical review of PET Experiments of Phonological Processing


Positron emission tomography (PET) is an experimental procedure that can be used in the investigation of the functional anatomy of auditory and phonological processing. The process often begins with setting of stimulus which are designed in determination of the brain that are significantly during speech and non-speech acoustic processing particularly for stimuli with or without promptly changing auditory cues (Fiez et al 357). The performance of auditory target detection tasks using the set stimuli produced higher activation in superior temporal, medial frontal cortices and the frontal opercula. These findings are however relative to the visual fixation control task. Poeppel (1996, p. 5) raises numerous concerns about the methodologies and the reported results for about eight studies concerning phonological processing emanation from six different neuroimaging laboratories. The main objective of this paper is to provide a critical response to Poeppel’s understanding and criticisms of PET experimentations of phonological processing.

General commentary

Inasmuch as the analysis by Poeppel (318) fails in the provision of a new understanding into the iconfines of PET methodology especially in the ways by which future neuroimaging services could be improved, his criticism largely emanate from failure of understanding. One way by which this failure is demonstrated is through the inability of Poeppel (319) to understand some of fundamental issues that motivate studies dedicated to functional imaging. Poeppel offers a challenge to experts in neuroimaging to base their experiments on the intension to ensure some form of clarity in their experimental designs, the analysis of the findings of the experiment and interpretation of these results in relation to the main objective of any PET studies that focus on phonological processing (Fiez et al 358). Poeppel (322) holds the belief that PET results derived from different experiments do not converge as expected especially in their ability to help in the production of a common understanding about functional imaging. This is derived from the understanding that similar experiments, which are designed in an attempt to separate similar language processes, reveal some form of stimulation in non-intersecting cortical expanses.

According to Poeppel (317) there is no region or a set of regions that were found to be ominously dynamic in all the studies that were conducted. This assertion by Poeppel raises questions concerning the possibility of the results converging to identify an area or a set of areas which were equally active in all the experiments focusing on phonological processing. A possible objection to this understanding is that it may be relatively impossible to hold such an expectation. This is because of the understanding that it is not possible to consider phonological processing as a unitary functioning initiative. In addition, it would also be prudent to note that the dissimilarities that exist between tsk comparisons such as the particular control conditions and presentation parameters used in developing comparisons must be considered throughout the study and the presentation of the findings.

It is factual that the experiment that Poeppel (338) chose to assess was relatively different and that these experiments, due to their dissimilarities did not have the capacity of isolating the same set of language processes. Inasmuch as it may be necessary to consider these factors in an attempt to validate the findings an assessing why there is incomplete overlapping between studies, it is also possible that there are relationships between some of the responsibilities involved in the experiments. This explains why it is possible to find out that there are different areas of activation that are in more than one experiment or task condition (Fiez et al 359).

Are there similarities in the revered experiments?

The assessments that Poeppel (318) makes in an attempt to identify the similarities between different studies is based on the conclusion that the studies appraised provide rhyme judgments as a way of engaging phonological processing (Poeppel 317). Poeppel’s arrival to this conclusion is an indication that numerous major issues have been ignored. An important issue is that not all rhyme judgments can be perceived as correspondent. This is despite the understanding that in all the experimental studies that Poeppel reviewed, participants in the studies were asked to make judgments concerning the sounds of different stimuli. One of the outstanding features that despite the similarities in the variables used in the studies, the said studies were in no way identical. An obvious distinction concerning these experiments was the type of segmentation tha the subjects were required to perform to ensure that they made appropriate sound judgments (Awh et al 160).

The outcome was that there were tasks that made it necessary for the motivations to be segmented into syllables. Other stimuli were segmented into either constant cluster or the remainder of the syllabus or others into individual phonemes. This can be perceived as the wealth of empirical evidence since phonemes, consonant cluster and syllables were a representation of fundamentally different phonological units whose differences could be found in factors such as their acoustic properties, the differences in the ages in which children could manipulate the phonological units and their relations to orthographic units. This explains why it is relatively less surprising when comprehensive philological and computational prototypes of speech production, word recognition and speech reception make explicit distinctions between these units (Awh et al 100).

There are additional differences emerging from the task conditions that were subjected to detailed reviews by Poeppel (318). The differences noted were in the short-term retention areas imposed by the units, the modalities of presentation, and the degree of stimulus exhibition from one item every three seconds. Furthermore, the type of stimulus process items in the form of word, letters, and non-words. The other difference was in the number of times that an item was presented to each of the participants in the study. This fluctuated from one time on every subject and numerous times per scan. The essence of these differences is based on the understanding that they are crucial in the determination of the eventual results of every study (Awh et al 100). For example it is possible to realize that the existence of any differences in the type of dispensation used in the realization of a phonological task may be varied. This is highly dependent on whether stimuli used are visual or auditory. Prior experiments and stimulus presentation parameters can have an effect on not only the subsequent performance of every task that employs the use of the item but also that which affects the response of blood flow to the items (Awh et al 101).

Failure by Poeppel (322) to consider the paramount essence of the standard circumstances in every task experimented that he critically reviewed can also be cited as reason for the profound dissimilarities in the results of the studies. This is based on the understanding that the underlying variances in the control conditions that are used in every experiments accounts for some of the non-overlapping activation patterns that are witnessed in these studies. For example, in the phonetic-passive language assessment, there seems to be no reason for the expectation of an stimulation of the superior temporal gyrus. This is because it is possible for the region to acquire activation from both conditions. The situation is however different in other studies examined by Poeppel. This is because these studies provide a situation where activation in this region can be anticipated considering the fact that there are no acoustic responses in the reference conditions. Furthermore, the decision by Poeppel (340) to compare phonetic conditions to noise conditions that lacked any language related auditory features; there was activation in the superior temporal gyrus and the Broca’s area.  

One of the major areas of concern for Poeppel (341) was the exploration for the region which is crucial to a specific, essential phonological function that was identifiable in every study. According to Poeppel (340), it should be possible to implicate a specific area or a set of areas in phonological processing. This is because some basic elementary computation necessary for phonological processing must have some consistency in the nature of their activity if phonological processing is entreated by the tasks. It is only later in the critical review that Poeppel acknowledged the possibility that phonological processing is not a unitary physiological operation (Poeppel 322). At this point Poeppel (341) develops a conclusion on the possibility that the entire process has been unacknowledged by PET researchers hence the possibility of a charitable interpretation of the no-overlap findings. Contrary to his assertions, PET researchers have acknowledged the view that phonological processing cannot be a unitary psychological process.

According to these researchers, cognitive tasks involve a convoluted variety of mental operations. This is because of the existence of the possibility of a multicomponent model that plays the role of an articulatory loop (De´monet 25). This model provides the suggestion that phonological processing can be realized through a network. Networks are derived from the phonemic constituent of the inner speech which in itself is a complex sensation. The existence of varieties of components can also be used in providing the understanding that there exists some independence in the phonological ideas and productivity encryptions that define the results of any cognitive processing procedures (Awh et al 106).

            It is more uncharitable to develop an assumption that phonological processing is not unitary, instead it is more rational. In a broader sense, it is possible to characterize phonological processes as those associated with the sensitivity of the spoken word. Furthermore, phonological processes may also involve transformations that are associated to the ways by which visual and auditory words are spoken or sound (Awh et al 106). This also involves developing an understanding that phonological processes may include the generation of sound based codes that are necessary in the production of words. The differences that exist in the transformations involve numerous processes that can be distinguished from the theoretical and empirical perspectives. For example, it is possible to associate some processes as acoustically based and others as articulatory based (De´monet 26).

There are times when the distinction between that acoustic and articulatory coding may not be elaborate enough. This is because of the differences in the levels of computing the phonological analysis. The computation process ranges from the individual phoneme and the level of the whole word. Inasmuch as all the experiments analyzed by Poeppel (330) engaged phonological tasks, the types of phonological processes that were required to perform the tasks were varied in each task. While operating on this understanding, it would be relatively surprising for find out that very study concerned with phonological processing produced almost similar degrees of activation in similar conditions.

The understandings that phonological tasks do not comprise unitary psychosomatic procedures have been addressed through past PET studies. These studies have been able to demonstrate that the variants that exist in a phoneme monitoring task produce dissimilar activation patterns (Awh et al 107). This proposes that under the power of certain investigational constrictions. The constraints include phonetic abstruseness and consecutive processing; varieties of cognitive strategies and neural structures can act as solutions to these tasks. The differences in the outcomes of the activation codes of the phonological procedures can be used as ways of developing alternative solution for understanding different abilities among individuals to cognitively engage in matters related to speech recognition and understanding (De´monet 27).

Can activation be found in non-overlapping areas?

Inasmuch as each experimental task may be exceptional, it does not possible necessitate the argument that all the phonological processes convoluted in the performance of each task were also unique to every study. Poeppel (332) notes the possibility of similarities in different tasks. This means that it is possible to develop an explanation that different there may be similarities between different tasks. In return, it would also mean that some areas of activation should be in numerous experiments.

Any claim that there exists a sparse or no overlap must be perceived as a misstatement of facts. A better understanding of this aspect is possible through a consideration of the position of stimulation described in the left inferior frontal gyrus in the experiments that were studied by Poeppel. This would mean that an anatomical coherence across varieties of experiments can be considered as an achievement due to the results of various experiments that were performed through varieties of PET scanners. These scanners had different ways of processing sonograms and spatial screen resolutions. The experiments of anatomical coherence also involved different methods through which radioactive water could be delivered (De´monet 27). An additional reason why the development of an argument that non-overlapping was a possibility in phonological processing must also be considered in relation to the sample size which varies considerably. Different individuals have varied cognitive abilities and this would mean that the outcomes of their findings must possess some similarities and differences (Awh et al 107).

            Through findings from different studies related to PET, it is possible to argue that there has been convergence in terms of the activations found in the Broca’s area. Other than the similarities, it is also possible to argue that there has been an overall conjunction in the understanding that is placed on activation (De´monet 26). Most of the researchers have argued that the area may be involved in different systems of subvocal articulatory depiction. This is because the existing theories provide an indication that the Broca’s area is essential in the functioning of the subvocal rehearsal system.

            In the process of developing the phonetic judgments, the participants must be provided with access to an articulatory representation which involves neural circuits within the Broca’s area (De´monet 26). It may be possible that the activation in the Broca’s area has an association with the phonemic constituent of interior speech which is a revelation of a complex phenomenon. The Broca’s area may also be involved in the retrieval of internal representation concerning the way a motivation resonances. It leads to the development of an understanding that internal representation may be founded partly on high level articulatory coding. However, it is important for additional research to be conducted n the subject matter to enable to development of a conclusive argument concerning the role of Broca’s area besides  providing answers to questions related to whether the said area contains functional sub regions (Awh et al 108).

This discussion together with those form additional studies on PET provides a conclusion that is different from that of Poeppel: converging activation in different experiments can be found and progress in developing an understanding the purposeful localization of the areas involved in phonological processing. There are different studies that have also been able to concur with People’s argument that there exists genuine non-overlap in different data especially that which focuses on inferior temporal gyrus stimulation, and the right hemisphere. These areas are often involved in short term memory tasks. Non-overlapping data arising from experimental noise can be perceived as false positives for a statistical understanding (Awh et al 110).

Possible criticism about PET experimental design and analysis

Poeppel provides an analysis of the factors that he considers to be the major cause of problems to the possibility of developing objective results whenever an individual is involved in PET experiments and analysis. The sparse overlap in the view of Poeppel (322) can be associated with ineffectively detailed task-control matching and task decomposition. In addition, it is possible for the sparse overlap to result from insufficient contact the linguistic theory, cognitive psychology and psycholinguistics. These are schools of thought that are crucial to the development of an elaborate understanding on matters related to the effective use of language in the determination of the best approach to the study that can generate sufficient and objective results (De´monet 26). The use of subtractive PET methodology in the study of neural representation and language processing may generate inherent problems throughout a study. It is possible to agree tha the factors that generate sparse overlap are integral in the development of an experiment design in the interpretation of PET studies. However, the analysis developed by Poeppel concerning the challenges they present in relation to the objectivity of the findings can in most cases be perceived as superficial and incorrect (De´monet 27).

The process of evaluating experimental designs of Poeppel’s criticisms generates an understanding that his criticisms arise from the conviction that the main goal of an experiment is to isolate a specific cognitive computation (320). Thus a discussion on how every charge can be disintegrated into a plethora of subcomponent leads Poeppel (344) to the conclusion that for every task comparison under investigation, it was not possible to isolate phonological process.

Nevertheless, the suggestion that there exists elementary cognitive operations and that it is possible to localize these operations does not necessitate the meaning that PET researchers operate on the understanding that it is only a single task comparison that can isolate any of the elementary operations (Awh et al 110). Various researchers have emphasized that elementary operations for an intricate purpose such as phonology have been partly understood. This is based on the possibility that it is relatively impossible to engage in designing of tasks and their control pair which are in contrast with one fundamental operation.

In addition, it is also based on the understanding that any language task has the ability to activate a set of distributed regions. The possibility that a charge can be disintegrated into numerous sub-components means tha t there are processes that may occur despite their irrelevance in the efficacious enactment of a task (Awh et al 111). The understanding that there exist differences on how different individuals perform their talks can be used in developing arguments that the decomposition of different operations facilitates the process of determining the subcomponents that are more prevalent and dominant in different individuals. Through such an argument it can be possible to assert that the differences in the ability of different persons to understand different aspects of language emanates from differences in their cognitive abilities (De´monet 26).

The understanding that one subtractive evaluation cannot be used in the segregation of a specific language operation helps in the generation of the argument that it is always important to consider the results from other methodologies and tasks comparison. Poeppel (333) ignores the ability of such converging evidence to provide a sufficient explanation in the differences involved in PET experiments with reference to the phonological process. The failure by PET researchers in the view of Poeppel (346) is based on the assumption that these individuals have had insufficient contract with disciplines that are essential in understanding the relevance of the phonological process.

Disciplines such as cognitive psychology, linguistic theory and psycholinguistics are essential in facilitating the development of a broader scope on matters related to the phonologic process. These experiments in the view of Poeppel do not seek concepts that go way beyond the basic understanding of the phonological process. The basic concepts that are considered in the view of Poeppel are not heavily founded on any hypothetical structure or model and there exists only limited reference to the theory of psycholinguistics which does the role of investigating speech (Poeppel 321).  The statement as presented by Poeppel ignores the assumption some of the studies were conducted in collaboration with subjected who had a broad understanding of the requirements of the studies.

There is absence of clarity on the side of Poeppel (334) on what he means by inadequate interaction with other disciplines and the disciplines that he considers to be worth considering prior to the organization of PET experiments. Inasmuch as he acknowledges tha there are numerous and self-sufficient models which can be essential in the development of a broad understanding of phonological processing, he fails to provide an elaborate perspective concerning the requirement of these models of thought. He does not assess the extensive theoretical analysis which in most cases produces a se4ries of divergent results.

In many situations, some of these models are never meant to provide an elaborate evaluation of the specific constituents of a specific task. Instead, they provide a framework that can be used in understanding the complex processes involved in PET experiments. This explains why in the process of designing a paradigm, it is expected that the PET investigators will derive their motivation from an ideal and still have to be engaged in process of selecting specific tasks. The process of selecting these tasks has its basis on experimental evidence that often includes different aspects of phonological processing. Furthermore, this process also requires that the phonological requirements of these tasks are different from other tasks and the control processes developed in the study.

The traditions of experimental psychology have in most cases been used by most imaging studies which rely on tasks in which the stimuli and the responses of the subjects can be constrained and defined. Poeppel (334) questions the ecological validity of task in PET experiments imply in different ways that the invalidity of the tasks can be derived from the understanding that they are unusual. Tachistoscopic presentation of any two virtual items in the view of Poeppel does not comprise a typical language processing computation. This is because the listeners are not often engaged in a typical process of breaking down the speech tributary into duos of the CVC strings that are in most cases considered to be temporarily adjacent. The argument that seeks to falsify the assumptions developed by Poeppel (335) is based on the understanding that any experiments task is a mimic of real life events.

Therefore, any results generated from these experiments, especially those that focus on the cognitive function of human beings must be considerd in relation to the understanding of real life events. It is a fact that those responsibilities are simulated, but whether the use of unnatural tasks may signify the generation of inconclusive results is dependent on the aim of the experimenter. The experimenter must consider whether his or her objective is to conduct controlee experiments of to produce conditions that can be perceived as natural (De´monet 28).

From a general perspective it is impossible for experimenters to realize both conditions since scientific experiments require the control of specific variables and the systematic manipulation of other variables in ways that produce the intended results. The understanding that science can either be engaged in artificial or natural procedures in an exclusive sense means that the intended results in PET experiments cannot be acquired outside the laboratory (Fiez et al 368). Experiments in the psychological discourse possess a plethora of inexplicable tasks that have confirmed to be essential despite lacking in ecological validity. In normal situations rats do not swim in pools of milk and human beings do not engage themselves in the activities such as looking at upside-down letters and making decisions on whether these letters represent mirrored images?

These activities are systematically manipulated to assess the relevance of phonological processing in understanding different linguistic aspects about man. In addition, they also constitute a platform that can be used in the development of effective measures that can be used in the development of strategies on how man can be provided with biter ways of speech mastery. In addition, these experimental procedures are sometimes ,manipulated to make it possible for PET researchers to be engaged in a process of developing paradigms for exploring specific mental processes that involve spatial memory and visual memory (Awh et al 122).  

Throughout his critical view, Poeppel (341) fails to provide a constructive remark or suggestion on the best ways that researchers should implement in improving on the experimental paradigms. He only enhances in a continuous process of raising criticisms on past studies by recognizing their failures especially on how the PET researchers were not motivated by psycholinguistic models. In addition, he does not provide a specific way that can be used in construction elaborate and objective experiments. From this criticism, it is possible to argue that the process of criticizing already existing experiments might be easy since it only involves the recognition of mistakes.

However, it is far more complicated to develop investigational task blends that can be practical in terms of their implementation and that which can meet the criteria that Poeppel (343) identifies as effective task disintegration control matching of tasks. Through such a process it would be easier to meet Poeppel’s objective of using natural language tasks and meeting the general agreements concerning that which is being measured. It is possible to assert that the arguments presented by Poeppel considered the artificial representation of any experiment as failing in the objective of relating tests to that which is considered a natural process.

Intrinsic complications in  the use of subtractive PET methodology

In the process of discussing the demerits of PET research methodology, Poeppel (334) argues that the differences in experimental task parameters such as the rate of presentation may produce inconsistencies in the findings. This does not mean that the process of analysis different experimental parameters in various PET researchers have been the source of variability in phonological process studies. In addition, Poeppel (336) also develops an assumption that most multiple trails in PET researches are performed during scans. This is a factual statement. This is because while a difference image is a representation of the mean change across manifold trails, it is relatively impossible to see why this can be perceived as a weakness of PET studies. Majority of studies in cognitive psychology employ the use of statistical variable reaction times as platforms for the evaluation of cognitive processes. This would lead through the development of an argument that the significant contribution of statistical variations such as means and median in understanding cognitive process has been the sole foundation of process that has been made in PET experiments (Awh et al 198).

            Poeppel (337) in an attempt to critic the role that PET plays in the development of a deeper understanding of phonological processing argues that it is relatively difficult for cognitive processes to add in a manner perceived as hierarchical. This assumption has been considered relatively weak in understanding the operations of PET experiments. This is because unlike the studies that focus on reaction times, the violations of the assumptions presented by Poeppel do not make results emanating from PET experiments inconclusive or uninterruptable. Successful comparison across multiple levels in any hierarchical design, especially in areas that are in violation of the categorized expectations can in most cases be identified directly.

            The ability of investigators to recognize this assumption has facilitated their ability to engage in the comparison of task conditions and numerous control conditions. There are also claims by Poeppel (338) that subtraction is an in-built assumption of a forward only phonological processing. This is however contrary to the normal operations of phonological processing because comparison images often portray dissimilarities between two conditions. There exists no fundamental assumption on the foundations of neuronal accomplishments which are responsible for the production of the differences (De´monet 27). Therefore the absence of initial reason for developing an assumption that only vertical effects can be found in the difference images of PET researches. Contrary to the assumptions of Poeppel, any experimental designs directed towards the isolation of top-down effects have been able to achieve their objectives.

            There are additional arguments by Poeppel (339) which are founded on the assumption that it is possible for subjects to engage in an automatic activation of representations that are irrelevant to the task in hand. There are however several strategies that have been developed to ensure that the process of evaluating and overcoming the impact of such subsidiary activation has already been developed (De´monet 27). For example, in the contemporary PET experiments, comparisons have been made while using control conditions that do not necessitate the possibility invoking automatic lexico-sematic processing. In addition, through PET research it has become possible for manifold task assessments have been used in the evaluation of whether it is possible to account for changes through such incidental activation. Tasks have also been developed to assess the possibility understanding undeviating consideration to different features of the same linguistic input (Awh et al 200).   

            One of the main sources of controversy between mind-related scientist and specialists in brain related matters such as neurologists is the absence of an obvious relationship between theoretical assumptions developed by the linguists and psychologists and the objects used in the description of these theories by the neurologists. Scholars in the mid related studies such as Sigmund Freud recognized the fact that it is possible the brain correlated of the functions of language may be comprised of a vast and continuous domain which spreads over the left hemisphere perisylvian cortex. The phonological process in this view is therefore facilitated by a network of interconnected regions with subserves that functions of language that may be impaired by lesions in this region (Fiez et al 372).

Different lesion sites have also been associated with various phonemic disorders. This does not mean that the complex aphasiological data, though difficult to interpret, is irrelevant. This assumption is contrary to the understanding of Poeppel who argued that different lesions, which are accountable for the existence of specific aphasic symptoms, make it relatively impossible to rely on the data derived from aphasiological studies. The contradictions found in the arguments by Poeppel (321) can be used in the development of an understanding that functional imaging techniques have a heuristic value for cognitive neuroscientific practices since they embrace a new experimental paradigm, the cognitive activation, in the production of results that are free from the assumptions developed by previously unique lesion based framework. Any evidence derived from activation studies can be considered as essential in the contribution of a refined neurocognitive model (Fiez et al 372).      

It is relatively reasonable to develop an assumption that the same type and approach of the brain-language isomorphism may prevail within the cognitive activation model as in the case of the aphasiological approach. That is the baring counterpart of language functioning constitute the activated networks that are distributed over the cortex instead of the isolated foci of activation. This has been perceived as one of the greatest drawbacks of functional imaging to individuals without sufficient knowledge on how PET experimentation operates. The process of analyzing and interpreting results is basically founded on a subtractive methodology which seeks an increase in the flow of blood in an experimental condition when another task is considered as a point of reference (Fiez et al 370). The fundamental parallelism of language processing and the automatic engagement of a variety of these language tasks cannot make it possible for a single comparison to be used in isolating any language operation. A combination of several comparisons and isolation mechanism permits the possibility of identifying the similarities and differences that are inherent in the development of an effective understanding of the phonological process (Fiez et al 370).


While this paper disagrees with most of the conclusions propounded by Poeppel (1996), it is possible to argue tha there is need for additional research on fictional imaging experiment to realize the development of a comprehensive account on the essence of neuroanatomical essence of phonological processing. It would be unrealistic to develop any form of understanding and conclusive assertions using eight experimental studies whose differences are in a span of six years. Different scientists have been involved in a process of decades of studies. Prior to the development of function imaging, there have been disagreements concerning the nature of the phonological processing among linguists and psychologists. The disagreements were associated with the differences in the process of interpreting the phonological processes involving visual word recognition.

Works Cited

Awh, E., Jonides, J., Smith, E. E., Schumacker, E. H., Koeppe, R., & Katz, S. Dissociation of

storage and rehearsal in verbal working memory: Evidence from PET. Psychological Science, in press.1996

De´monet, J.-F., Price, C., Wise, R., & Frackowiak, R. S. J. 1994b. Differential activations of

right and left posterior sylvian regions by semantic and phonological tasks: a positron emission tomography study in normal human subjects. Neuroscience Letters, 182, 25– 28

Fiez, J.A., Raichle M, E., Miezin, F. M & Katz, W. F. PET Studies of Auditory and Phonological

Processing: Effects of Stimulus Characteristics and Task Demands. J Cogn Neurosci. 1995 Summer;7(3):357-75. doi: 10.1162/jocn.1995.7.3.357.

Poeppel, David. A critical review of PET studies of phonological processing. Brain and

Language, in press. 1996