A panel of three neuroscientists classified all language comprehension and nonverbal tasks administered to linguistic isolates according to the minimal neurobiological requirements necessary to obtain the correct answer (Vyshedskiy et al. 2017). Specifically, panel members analyzed (1) The minimum number of disparate objects that needed to be purposefully imagined together in order to answer correctly, called the number of objects (NOB) score, and (2) The type of object modification required by the question (see details below). Members of the panel analyzed each question independently of each other and then discussed questions one by one to reach a unanimous opinion.

(2.1) Establishment of the type of object modification: For any object in the mind’s eye, we can voluntarily change its color, size, position in space or rotation. The mechanisms of these processes involve PFC-controlled modification of the object’s representation in the posterior cortex. As related to the posterior cortex neuronal territory, such mechanisms can be classified into three classes: 1) those that do not involve any object modification; 2) those that involve modification of activity in only the ventral visual cortex (Fig. 1); and 3) those that involve coordination of activity in both the dorsal and ventral visual cortices. For example, object comparisons does not involve modification of any object; color or size modification of an object recalled in memory is limited to the ventral visual cortex (Gabay et al. 2016); finally, location in space modification or rotation involve coordination of activity in both the ventral and dorsal visual cortices (Cohen et al. 1996, Goodale and Milner 1992, Lee et al. 2006, Schendan and Stern 2007, Zacks 2008). Accordingly, the three types of questions were assigned the posterior cortex territory score (PCT score) ranging from zero to two, with zero corresponding to minimal posterior cortex territory and two corresponding to greatest amount of posterior cortex territory.

1. Find the named object. A common task requires the subject to find a named object from a variety of physical objects (or pictures) to choose from. These tasks are based of subject’s ability to compare the internal representation of the target object with all the possible solution objects, primarily relying on matching from memory. Since this type of question does not require any mental combination of disparate objects, it is assigned a NOB score of one. Furthermore, since this type of question does not require any modification of the object, it corresponds to the PCT score of zero (i.e. requiring the least amount of the posterior cortex territory).

2. Integration of modifiers in a single object. Another common comprehension test requires the subject to integrate a noun and an adjective. A subject may be asked to point out the picture with a {yellow/red/green} + circle placed among several decoy images with {yellow/red/green} + {triangle/square} thus forcing the integration of color and noun. Similarly, to integrate size and noun one may be asked to point to a {big/small} + {circle/triangle/square}. Neurologically, the integration of modifiers involves the modification of neurons encoding a single object and therefore has a NOB score of one. As color and/or size were modified, these tasks were given a PCT score of one. (modification is limited to the ventral visual cortex, see Gabay et al. (2016)).

2.1. Comparative. In a comparative test the subject may be shown two circles of different sizes and asked to point out the circle that is {bigger/smaller}. In terms of visual representation in the posterior cortex, the circles can be considered one at a time, making the comparative task a special case of integration of the size modifier task. Therefore, minimally, the comparative task involves the modification of neurons encoding a single object (the circle) and has a NOB score of one. In terms of the PCT, these questions were assigned a score of one (modification is limited to the category of size represented in the ventral visual cortex, see Gabay et al. (2016)).

2.2. Superlative. In a superlative test, the subjects may be shown several circles of different sizes not aligned by size and asked to point out the biggest/smallest circle. Similar to comparative task, superlative task is a special case of integration of the size modifier, it involves the modification of neurons encoding a single object and therefore has a NOB score of one. In terms of the PCT, these questions were assigned a score of one (modification is limited to the ventral visual cortex, see Gabay et al. (2016)).

3. Integration of number modifier. To integrate the number and noun the subject may be asked to point out a card with {two/three/four} circles among distractor cards with {two/three/four} + {squares/triangles}. Similar to comparative task, the integration of number modifier minimally involves the modification of neurons encoding the number: independently of whether the number is two or a million, only a single object (the circle) is visualized and therefore this test has a NOB score of one. Reports indicate that numerical information is represented by regions of the posterior parietal lobes (Dehaene et al. 2004, Nieder and Dehaene 2009). Thus, these tasks are given a PCT score of two.

3.1. Singular vs. plural. Singular vs. plural tasks involve showing several pictures and asking the subject to point out the correct picture: ‘Show me the flower’ versus ‘Show me the flowers. Neurologically, in terms of visual representation of object in the posterior cortex, these tasks are the special case of the integration of number modifier; they involve the modification of neurons encoding a single object (NOB score of one) and have a PCT score of two.

3.2. Negative vs. affirmative statements. To test understanding of simple negation the subject is shown a pair of pictures identical except for the presence or absence of some element. For example, the subject may be asked to indicate a picture in which the rabbit has a carrot vs. the one in which the rabbit does not have a carrot. Neurologically, in terms of visual representation of object in the posterior cortex, these tasks involve the modification of neurons encoding a single object (e.g., a carrot) and therefore have a NOB score of one. These questions are a special case of integration of number modifier, with the number set to zero. Accordingly, they were assigned a PCT score of two.

4. Mental rotation and modification of a single object’s location in space. A number of verbal and nonverbal tasks given to linguistic isolates require mental rotation or other modification of an object’s location in space. An example of a nonverbal mental rotation task is presented in Fig. 2. Mental rotation and modification of a single object’s location in space involve the PFC-coordinated activity in both the ventral and the dorsal visual cortices (Cohen et al. 1996, Harris et al. 2000, Schendan and Stern 2007), and are therefore assigned the PCT score of two. However, these tasks are still limited to a single object and, consequently, assigned a NOB score of one.

4.1 Copying stick and block structures. Copying stick and block structures test is a nonverbal test that relies on mental rotation of sticks or block pieces in order to align them with the instructions. Since mental manipulations involve one piece at a time, these tasks are limited to a single object and, consequently, assigned a NOB score of one. Mental rotation involve the PFC-coordinated activity in both the ventral and the dorsal visual cortices (Cohen et al. 1996, Harris et al. 2000, Schendan and Stern 2007), and are therefore assigned the PCT score of two.

5. Mental synthesis of several objects.

5.1 Spatial prepositions. Understanding of spatial prepositions such as in, on, under, over, beside, in front of, behind requires a subject to superimpose several objects. For example, the request “to put a green box {inside/behind/on top of} the blue box” requires an initial mental simulation of the scene, only after which is it possible to correctly arrange the physical objects. An inability to produce a novel mental image of the green box {inside/behind/on top of} the blue box would lead to the use of trial-and-error, which in majority of cases will result in an incorrect arrangement. Understanding an instruction to “put the bowl {behind/in front of/on/under} the cup requires the combination of two objects and therefore is assigned a NOB score of two. Understanding an instruction to “put the cup in the bowl and on the table” requires the combination of three objects and therefore is assigned a NOB score of three. In general, a NOB score in mental synthesis questions is defined as the number of disparate objects that have to be imagined together.

When two disparate objects are imagined together their spatial relationship has to be encoded. Encoding of spatial information is the function of the dorsal visual cortex (Cohen et al. 1996, Harris et al. 2000, Schendan and Stern 2007). Accordingly, mental synthesis of multiple objects always involves the PFC-coordinated activity in both the ventral and the dorsal visual cortices and is therefore assigned the PCT score of two.

5.2 Flexible syntax. The order of words, or syntax (from Greek syn, meaning together, and taxis, meaning an ordering), is an essential quality of all human languages. A change in the word-order often completely changes the meaning of a sentence. For example, the phrase “a cat ate a mouse” and the phrase “a mouse ate a cat,” have very different connotations only because the word order was changed. To understand the exact meaning of these two sentences, the subject must mentally synthesize the objects of “a cat” and “a mouse.”

Flexible syntax is distinguished from rigid syntax, whereby the meaning of a word is always dictated by its position in a sentence. For example, if we all agree that the first noun in the “noun1 ate noun2” sentence will always be the eater and the second noun will always be the prey, then the meaning of the sentence can be inferred without mental synthesis of multiple objects. When you hear “a cat ate a mouse,” you automatically know that noun1 (a cat) was the eater and noun2 (a mouse) was the prey. You do not need to synthesize the image of a cat and a mouse together; you can simply infer the meaning from the sequence of words in the sentence. Of course this rule constricts the conversation to a single sentence structure: “noun1 ate noun2.” If you follow this structure strictly, you will always misinterpret sentences like “a cat was eaten by a mouse,” “a cat will be eaten by a mouse,” “this cat never eats mice,” since you will always think that noun1 (the cat) is the eater. Human languages are not bound by limitations of rigid syntax: both the first and the second nouns can be the ‘eaters’ (“a cat ate a mouse” versus “a cat was eaten by a mouse”). All human languages use flexible syntax and rely on mental synthesis of multiple objects to decode the meaning of a sentence.

Similar to spatial prepositions tasks, a NOB score in flexible syntax tasks is defined as the number of disparate objects that have to be imagined together and PCT score of two is assigned.

5.3 Before/after word order. To test understanding of before/after word order, subjects can be asked ‘to touch your nose after you touch your head.’ Similar to flexible syntax, this task relies on mental synthesis to simulate the process of touching the nose and the head. These tasks are given a NOB score of two and PCT score of two.

5.4 Active vs. passive. Subjects can be asked to point to a picture showing ‘The boy is pulling the girl’ vs. ‘The girl is pulled by the boy.’ Similar to flexible syntax, this task relies on mental synthesis to combine the boy and the girl in the mind’s eye. It is given a NOB score of two and PCT score of two.

5.5 Exact arithmetic abilities. Mentally finding the solution to a complex arithmetic problem such as mental two-digit addition (e.g. 43+56) and multiplication (e.g. 32x24) also relies on mental synthesis. This process is different from multiplication of single digit numbers (e.g. 2x4) that can be retrieving from memorized multiplication table (Zago et al. 2001). Since the exact number of objects in two-digit multiplication is not easily assertained, this task received an NOB score of “2+.” Mental two-digit addition and multiplication were assigned a PCT score of two.

6. Special cases

6.1. Classification tasks. Classification tasks involve sorting objects on the principles of gender, animacy, etc. Classification tasks involve no object combination and therefore were assigned a NOB score of one. The tasks also involve no modification of objects and therefore were assigned the PCT score of zero. Classification may be more challenging than similarly scored “Find the object” tasks, since the object’s class has to be determined, but this is not reflected in the scoring system used in this report.

6.2 Picture arrangement into a story. A typical test consists of three to six pictures with cartoon characters that can be arranged into a story. For example, picture A shows a boy running, picture B shows the boy falling, and picture C shows the boy crying. The pictures are presented to the subject in scrambled order with instructions to rearrange the pictures to make the best sense. A typical adult might approach this task by mentally animating the characters and generating a story using mental synthesis of multiple objects. However, in order to identify the minimal neurobiological requirements for the correct answer, we considered the following alternative. Most picture arrangement tasks rely on subject’s understanding of causality between events shown in the pictures. The inference of causality can be made from a memory of similar events watched on TV or a personal experience, e.g. running, falling and crying thereafter. The process of inferring the causality does not rely on combination or modification of objects and therefore was assigned a NOB score of one and the PCT score of zero.

The summary of NOB and PCT scores for all linguistic tests is provided in Table 1.

Fig. 3 summarizes performance of linguistic isolates in verbal and nonverbal tests. This section explains some of the calculations reported in Fig. 3.

Comments:

1. The methodology for deriving a NOB and PCT scores from the performance IQ score was described (Vyshedskiy et al. 2017).

2. Genie was tested with Raven’s Matrices and Wechsler Intelligence Scale for Children (WISC). Fromkin et al. (1974) report that 1.5 years into rehabilitation Genie’s “performance on the Raven Matrices could not be scored in the usual manner but corresponded to the 50th percentile of children aged 8.5 to 9 years”. Curtiss et al. (1975) report that 2 years into rehabilitation “on WISC... performance subtests ... she has achieved subtest scaled scores as high as eight and nine”. In addition, Curtiss (1977) reports that 5 years into rehabilitation “Genie was given the Coloured Progressive Matrices Test. ... Genie’s overall score was 29.”.

Using this limited information we attempted to map Genie’s IQ into a NOB and PCT scores:

A. WISC uses three sections to determine the performance score: Matrix Reasoning, Figure Weights, and Visual Puzzles. A person who can only answer test questions with a NOB score of one and PCT score of zero (i.e. “Find the same object” and “Amodal completion”) would receive scaled section scores of 2, 5, 3 respectively (Vyshedskiy et al. 2017). A person who can also answer test questions with a NOB score of one and PCT score of one (i.e. “Integration of size/color modifier”) would receive scaled section scores of 4, 5, and 4, respectively. A person who can also answer test questions with a NOB score of one and PCT score of two (i.e. “Integration of number modifier” and “Mental rotation and modification of object’s location in space”) would receive scaled section scores of 6, 10, and 4, respectively. To demonstrate ability for mental synthesis of multiple objects, the subject has to have section scores above 6, 10, and 4, respectively. From the Curtiss et al. (1975) report that Genie“ achieved subtest scaled scores as high as eight and nine” we have to assume that the highest scaled section score was “eight and nine”. The easiest WISC section where Genie would be expected to show the highest score is the Figure Weights sections. The score of “eight and nine” in the Figure Weights section corresponds to a NOB score of one and PCT score of two. That is Genie demonstrated ability to “Find the same object,” “Integrate of size/color modifier,” “Integrate of number modifier,” and “Mentally rotate and modify object’s location in space,” but failed questions that required mental synthesis of multiple objects.

B. From Fromkin et al. (1974) report that Genie’s “performance on the Raven Matrices... corresponded to the 50th percentile of children aged 8.5 to 9 years”, assuming that Raven Standard Progressive Matrices test was administered and consulting the Raven manual (Raven et al. 2000), Table SPM12, we gather that Genie’s raw score was between 33 and 36. This raw score corresponds to a NOB score of one and PCT score of two (Vyshedskiy et al. 2017) and confirms that Genie demonstrated ability to “Find the same object,” “Integrate of size/color modifier,” “Integrate of number modifier,” and “Mentally rotate and modify object’s location in space,” but failed questions that required mental synthesis of multiple objects.

C. The latest nonverbal test was administered 5 years into rehabilitation “Genie was given the Coloured Progressive Matrices Test ... Genie’s overall score was 29.” (Curtiss 1977) Raven’s Colored Progressive Matrices was originally designed for children aged 5 through 11 years-of-age, the elderly, and mentally and physically impaired individuals. This test contains sets A and B from the standard matrices, with a further set of 12 items inserted between the two, as set Ab. Most items are presented on a colored background to make the test visually stimulating for participants. The score of 29 means that Genie correctly answered 29 out of 36 questions. Assuming Genie has answered correctly the easiest question, we infer that she answered correctly all 12 questions in the set A (maximum NOB score of one and PCT score of zero), 6 questions in the set B (maximum NOB score of one and PCT score of one), and 11 questions in the set Ab (maximum NOB score of one and PCT score of zero) (Vyshedskiy et al. 2017). In the Coloured Progressive Matrices Test the hardest questions Genie was able to answer had a NOB score of one and the PCT score of one. As Genie was not able to answer any questions with a NOB score of two, we conclude that Genie failed those questions that required mental synthesis of multiple objects.

Thus, in all three nonverbal tests that Genie was administered up to five years into rehabilitation Genie has invariably failed to demonstrate capacity for mental synthesis of multiple objects.

3. Grimshaw reports that E.M.’s “performance on both simple and complex modification tasks was above chance but imperfect from the earliest assessment. E.M. made more errors in complex modification (two adjectives) than in simple modification (one adjective). Almost all errors involved a failure to identify the correct shape—the modifiers were always correct” (Grimshaw et al. 1998).

4. Curtiss reports that “She didn't perform as well as Genie on such tasks, but she could put picture sequences into a logical sequence” (personal communications).

5. Curtiss reports that “She did not appear to comprehend (or use) such elements” (personal communications).

6. Curtiss (2014) reports that “On standardized (nonverbal) psychological and neuropsychological tests Chelsea consistently demonstrates a performance IQ of between 77 and 89”. It is unclear what exact nonverbal IQ tests were used with Chelsea. For most nonverbal tests, the score of 78 corresponds to a NOB score of one and PCT score of one; the score of 86 corresponds to a NOB score of one and PCT score of two (Vyshedskiy et al. 2017). We conclude that at best Chelsea has demonstrated an ability to “Integrate number modifier,” and “Mentally rotate and modify object’s location in space,” but failed those questions that required mental synthesis of multiple objects.

7. Morford (2003) reports that “Maria was administered Raven’s Standard Progressive Matrices after three months and after 32 months exposure to (American Sign Language) ASL. Using norms gathered in 1986 for children in the US, her score after three months’ exposure to ASL was still below the fifth percentile for her age group, but after 32 months’ exposure to ASL, her performance was exactly at the fiftieth percentile for her age group”. Fifth percentile corresponds to the raw score of 35 and IQ score of 75 (Raven et al. 2000), a NOB score of one and PCT score of two (Vyshedskiy et al. 2017). Fiftieth percentile corresponds to the raw score of 52 and IQ score of 100 (Raven et al. 2000), a NOB score of two and PCT score of two (Vyshedskiy et al. 2017), indicating that Maria has correctly answered nearly 20 questions that required mental synthesis of multiple objects.

8. Morford (2003) reports that “Marcus was administered four of the WISC-R Performance Subtests after 19 and after 31 months exposure to ASL... Marcus’s performance improved by more than ten scale points from the first to the second administration, increasing the corresponding performance IQ from 70.5 to 85 points. The WISC score of 70.5 corresponds to a NOB score of one and PCT score of zero (Vyshedskiy et al. 2017). The WISC score of 85 corresponds to a NOB score of one and PCT score of between one and two. We conclude that at best Marcus has demonstrated and ability to “Integrate of number modifier,” and “Mentally rotate and modify object’s location in space,” but failed in questions that required mental synthesis of multiple objects.

9. Subject I.C. was unable to use spatial signs (left/right, on top of/underneath) to describe the relationship between the objects.

10. According to Ramírez et al. (2013), despite their age, the utterances of Shawna, Cody, and Carlos were neither long nor complex, never used lexical items indicating subordination or conditionals, and never used inflected verbs.

11. According to Ramírez et al. (2013) Cody received the score of 91 on the Test of Nonverbal Intelligence-3 (TONI-3) and the score of 85 on the Wechsler Nonverbal Scale of Ability (Wechsler 1949) indicating that he failed to answer most or all questions that required mental synthesis of multiple objects.

12. According to Ramírez et al. (2013) Carlos received the score of 85 on the TONI-3 and the score of 74 on the Wechsler Nonverbal Scale of Ability (Wechsler 1949) indicating that he failed to answer all questions that required mental synthesis of multiple objects.

Our visual world consists of meaningful, unified, and stable objects that move coherently as one piece. Objects, therefore, constitute the functional units of perception (Vallortigara 2004). Over time, neurons synchronously activated by visual observation of an object, get wired together by enhanced synapses and form a resonant system that tends to activate as a single unit resulting in perception of that object. These resonant networks of neurons encoding objects in the cerebral cortex are known as neuronal ensembles (Hebb 1949). When one perceives any object, the neurons of that object’s neuronal ensemble activate into synchronous resonant activity (Quiroga et al. 2008). The neuronal ensemble binding mechanism, based on the Hebbian principle “neurons that fire together, wire together,” came to be known as the binding-by-synchrony hypothesis (Singer and Gray 1995, Singer 2007).

While the term “neuronal ensemble” is often used in a broad sense to refer to any population of neurons involved in a particular neural computation, in this discussion we will use the term more narrowly to mean a stable group of neurons which are connected by enhanced synapses and which encode specific objects. Thus, neuronal ensembles are internal equivalents of objects: while the latter are physical or external units of perception, the former are the internal units of perception; each object that has been seen and remembered by a subject is encoded by a neuronal ensemble.

The neurons of an ensemble are distributed throughout the posterior cortex (occipital, temporal, and parietal lobes). Most neurons encoding a stationary object are located within the ventral visual cortex (also known as the "what” stream) (Goodale and Milner 1992, Fig. 1). These thousands of neurons encode the various characteristics of each object, such as shape, color, texture, etc. (Quiroga et al. 2009, Waydo et al. 2006). Most humans can use their PFC to purposefully change elements of a neuronal ensemble enabling an alteration of an object’s color, size, rotation, or position in space within the “mind’s eye.”

The PFC consists of two functionally distinct divisions. The ventromedial PFC is predominantly involved in emotional and social functions such as the control of impulse, mood, and empathy (Striedter 2004). The lateral PFC (LPFC) is predominantly, but not exclusively, involved in “time integrating and organizing functions, such as control of working memory and imagination” (Fuster 2008). The LPFC is evolutionarily related to the motor cortex and there are a number of pertinent parallels between the two (Striedter 2004, Fuster 2008). Just as the primary motor cortex can recruit motor units of voluntary muscles, the LPFC is able to activate neuronal ensembles in the posterior cortex. If motor units are homologous to neuronal ensembles and the motor cortex is homologous to the LPFC, then sequencing of muscle movements is homologous to sequencing of thoughts. The underlying mechanisms behind this sequencing of thoughts and their relationship to syntactic language are the primary focus of the following discussion.

We have previously hypothesized that the mechanism behind the mental synthesis of independent objects involves the LPFC acting in the temporal domain to synchronize the neuronal ensembles which encode those objects (Vyshedskiy 2014, Vyshedskiy and Dunn 2015). Once these neuronal ensembles are time-shifted by the LPFC to fire in-phase with one another, they are consciously experienced as a unified object or scene and could therefore be mentally examined as a cohesive unit. This hypothesis is known as Mental Synthesis theory (Vyshedskiy and Dunn 2015) and it is indirectly supported by several lines of experimental evidence (Hipp et al. 2011, Hirabayashi and Miyashita 2005, Rodriguez et al. 1999, Sehatpour et al. 2008, Uhlhaas and Singer 2006).

We have also hypothesized that the neurological process of modifier integration, whereby the LPFC modifies the activity of neurons in a single neuronal ensemble resulting in changes of an object’s perceived color or size, involves the LPFC acting in the temporal domain to synchronize the color encoding neurons in the visual area V4 with the object’s neuronal ensemble in the ventral visual cortex (Vyshedskiy et al. 2017). Once these neurons in V4 are time-shifted by the LPFC to fire in-phase with the object’s neuronal ensemble, the object is consciously experienced in the new color. LPFC-driven synchronization of neurons in the posterior cortex is likely a general mechanism necessary to achieve any internally-driven novel sensory experience. The LPFC can be viewed as a puppeteer and the objects encoded in the posterior cortex as its many puppets. By pulling the strings, the LPFC changes the firing phase of the retrieved neuronal ensembles, synchronizing them into new mental constructs. Thus, understanding of syntactic language requires the LPFC puppeteer to precisely schedule the firing phase of the neuronal ensembles puppets dispersed throughout the posterior cortex.

The Mental Synthesis theory also predicts that to synchronize neuronal ensembles dispersed throughout the posterior cortex, the LPFC must rely on synchronous connections to posterior cortex neurons (Vyshedskiy et al. 2017, Vyshedskiy 2014). In many animal models synchronous connections have been observed between multiple brain areas that depend on precise timing for communication despite varying path lengths. In the cat retina, axons from peripheral regions have a greater conduction velocity than axons from neurons at the center of the retina to assure the simultaneous arrival of impulses in the brain (Fregnac et al. 1987). Experiments in rats show that myelination is the primary factor producing uniform conduction time (to within 1ms) in connections between the inferior olive nucleus in the brainstem and the cerebellar cortex, despite a wide variation in axon length (Lang and Rosenbluth 2003, Sugihara et al. 1993). In cats, synchronous activation of groups of cells distributed in distant cortical locations has been shown in the visual cortex (Gray et al. 1989) and even between the two hemispheres of the brain (Engel and Konig 1991). It was also shown in rats that synchronous action potential delivery from the layer V pyramidal neurons in the ventral temporal lobe to target subcortical regions is based on the differential conduction velocity in each fiber (Chomiak et al. 2008), which appears to be best explained by differential myelination changing the conduction velocity of the individual axons (Kimura and Itami 2009). The synchronous connections achieved by differential myelination were also observed in cats between the amygdala and the perirhinal cortex (Pelletier and Paré 2002) and in mice between the thalamus and the somatosensory cortex (Salami et al. 2003). Accordingly, we hypothesized that synchronous frontoposterior connections achieved by differential myelination are essential for the LPFC’s ability to synchronize posterior cortex neurons located at various physical distances from the LPFC and that these connections are therefore essential for understanding of complex syntactic language, Fig. 4 (Vyshedskiy 2014).

Myelination is nearly completed by birth in most species in which the young are relatively mature and mobile from the moment of birth, such as wild mice (Tessitore and Brunjes 1988) and horses (Szalay 2000). However, in humans, myelination is delayed considerably. Few fibers are myelinated at birth and some brain regions continue myelination well into mid-life. The PFC region is the last region to complete myelination. Myelination of cortico-cortical fibers that connect the PFC to other cortical areas does not reach full maturation until the third decade of life or later (Bartzokis et al. 2001, Toga et al. 2006). To confer an evolutionary benefit, this delay in myelination must be associated with a practical advantage. Humans’ delayed maturation of myelination raises the possibility that experience determines myelination in a manner that optimizes synchronicity in neural networks.

There is significant evidence that neural activity can affect myelination. The number of myelin-forming cells in the visual cortex of rats increases by 30% when the rats are raised in an environment that is enriched by additional play objects and social interaction (Szeligo and Leblond 1977). Premature eye opening in neonatal rabbits increases myelin production presumably by increasing activity in the optic nerve (Tauber et al. 1980). In contrast, rearing mice in the dark reduces the number of myelinated axons in the optic nerve (Gyllensten and Malmfors 1963). The proliferation of myelin-forming precursor cells in the rat’s developing optic nerve depends on electrical activity in neighboring axons (Barres and Raff 1993). In neuronal cell culture, induction of action potentials with electrical stimulation affects axonal myelination (Stevens and Fields 2000, Stevens et al. 1998).

Environmental effects on axonal myelination extend beyond animal studies. In human infants, early experiences increase myelination in the frontal lobes in parallel with improved performance on cognitive tests (Als et al. 2004). On the contrary, children suffering from abuse and neglect show on average a 17% reduction in myelination of the corpus callosum, the structure comprised of connections between the neurons of the left and the right hemispheres of the brain (Teicher et al. 2004). Extensive piano practicing in childhood is accompanied by increased myelination of axons involved in musical performance, the white matter thickening increases proportionately to the duration of practice (Bengtsson et al. 2005).

To sum up, synchronicity in at least some neuronal networks seems to be achieved via differential myelination and myelination may be experience-dependent. In fact, considering the many variables affecting conduction delays in an adult brain and the hundreds of millions of frontoposterior connections, genetic instruction alone would seem inadequate to specify the optimal conduction velocity in every axon. Accordingly, the development of synchronous frontoposterior connections likely involves adjustment of the conduction velocity in individual fibers via an experience-dependent process.

The experience-dependent developmental mechanisms have been clearly demonstrated for the visual, auditory, somatosensory and motor systems in virtually all species, from humans to Drosophila (Barth et al. 1997, Berardi et al. 2000). Childhood sensory deprivation, such as monocular deprivation (Doupe and Kuhl 1999, Fagiolini et al. 1994, Harwerth et al. 1986, Huang et al. 1999, Issa et al. 1999, Olson and Freeman 1980) or monaural plugging (Knudsen et al. 1984) and alterations in sensory input, such as caused by strabismus (Berman and Murphy 1981), rearing in a restricted auditory environment, or misalignment of the auditory and visual space maps (Brainard and Knudsen 1998, King and Moore 1991), result in lifelong changes of neural networks and a deficit in the corresponding function. On the contrary, increased sensory experience has been associated with improved neural connections: in humans, musical training in childhood leads to an expanded auditory cortical representation and increased myelination of motor axons involved in musical performance (Bengtsson et al. 2005, Pantev et al. 1998) and introduction to a second language before the age of 12 results in a more optimal language encoding (Choi et al. 2017, Kim et al. 1997).

A common thread between these experience-dependent processes is that the ontogenetic experience essential for the formation of the neurological network is the same experience for which this network is naturally used in adulthood: stimulation of eyes by light is essential for formation of visual cortical connections, auditory stimulation is essential for formation of auditory cortex connections, motor practice is essential for formation of cortical motor connections, etc. Thus, it is compelling to hypothesize that cortical connections, which are essential for mental synthesis of multiple objects in adulthood are acquired as a result of the childhood mental synthesis exercises. Noting that conscious purposeful mental synthesis of disparate objects is primarily stimulated by syntactic language – whether it is internal language when we are thinking to ourselves or external language when we are involved in conversation – we would predict that the synchronous frontoposterior connections can only be acquired as a result of childhood exposure to syntactic language. The lack of syntactic language during the sensitive period of the PFC plasticity would result in asynchronous frontoposterior connections and mental synthesis disability.

In fact, this is exactly what we have observed in all linguistic isolates. Their lifelong mental synthesis disability is consistent with the critical role of early syntactic language use for establishment of synchronous frontoposterior connections. Childhood use of syntactic language (both externally and internally) likely provides the necessary training mechanism for development of synchronous connections between the LPFC and the sensory memory stored as neuronal ensembles in the posterior cortex. Without those synchronous connections having developed before the closure of the period of plasticity, the LPFC executive remains forever disabled at performing precise phase control of neurons, which are located over the large parts of the posterior cortex. The millions of strings connecting the LPFC puppeteer to its puppets (the neuronal ensembles in the posterior cortex) remain ill-adjusted, with the LPFC unable to control synchronization of independent neuronal ensembles and, therefore, unable to synthesize novel mental images. A child involved in a normal syntactic conversation (either with another person or with himself) internally manipulates mental images in his/her mind and during this process adjusts conduction velocity to achieve synchronicity in the neural networks connecting the LPFC to widespread regions of the posterior cortex.

The critical role of the syntactic language in formation of synchronous frontoposterior connections and mental synthesis is consistent with findings in late first-language learners and nonverbal individuals with autism. Elissa Newport and Rachel Mayberry studied the acquisition of sign language in deaf individuals differing in age of exposure: some were exposed to sign language from birth, while other children first learned sign language at school. These studies were conducted with adults who have been using sign language for at least twenty years. The results of the studies consistently indicated a negative correlation between the age of sign language acquisition and ultimate proficiency: those exposed to sign language from birth performed best, and late learners — worst, on all production and comprehension tests (Mayberry and Eichen 1991, Newport 1990). Furthermore, late first-language learners have syntactic difficulty describing events with complex temporal relations (Kocab et al. 2016), and have slower and less accurate performance in mental rotation independent of age and the total number of years of experience with the language (Emmorey et al. 1993, Martin 2009, Martin et al. 2013, Pyers et al. 2010). Emmorey et al. (1993) found that adults who had learned American Sign Language (ASL) early in life were more accurate than later learners at identifying whether two complex-shape figures presented at different degrees of rotation were identical or mirror images of each other. Martin (2009) found that men who learned ASL natively were faster than later learners at identifying whether two-dimensional body-shaped figures (bears with one paw raised) presented at different rotations were identical or mirror images of each other. Martin et al. (2013) demonstrated that even after decades of signing experience, the age at which one first learned Nicaraguan Sign Language was a better predictor of mental rotation accuracy than the total number of years of experience with the language. Across all ages, early learners were more accurate than late learners. Pyers et al. (2010) studied generational differences between the 2 cohorts of deaf Nicaraguan signers who learned the emerging sign language at the same age but during different time periods – the first cohort of signers who grew up when the sign language was in its initial stage of development with few spatial prepositions and the second cohort of children who grew up a decade later exposed to more spatial prepositions and other recursive elements. The study used two spatial tasks that required the use of a landmark to find a hidden object: in the disoriented search condition, participants entered a small, enclosed room with a single red wall as a landmark. They watched an experimenter hiding a token in one corner, and then were blindfolded and turned slowly until disoriented. Then the blindfold was removed and the subjects indicated the corner where they thought the token was hidden. In the rotated box condition, the procedure was similar: the token was hidden in a small-scale tabletop model of the room, the participant was blindfolded, and the model (not the participant) was turned. The second-cohort signers, now in their 20s outperformed the first cohort, now in their 30s, both when they were disoriented and when the box was rotated. The first-cohort participants made their decisions slowly and with effort, taking as long as nine seconds and reporting that the task was difficult. Even in neurotypical children, production of spatial terms from 14–46 months predicted their performance on nonlinguistic spatial tasks at 54 months (Pruden et al. 2011). Better and faster performance in spatial tasks in individuals exposed to more complex syntactic language earlier in life is best explained by the crucial role of syntactic language in fine-tuning the frontoposterior connections between the LPFC and the posterior cortex.

Prelingual deafness is such a serious concern that the US government has enacted laws to identify affected newborns. In 1999, the US congress passed the “Newborn and Infant Hearing Screening and Intervention Act,” which gives grants to help states create hearing screening programs for newborns. Otoacoustic Emissions Testing is usually done at birth, followed by an Auditory Brain Stem Response if the Otoacoustic Emissions test results indicated possible hearing loss. Such screening allows parents to expose deaf children to a formal sign language as early as possible and therefore avoid any delay in introduction to syntactic language.

Problems observed in late first-language learners also afflict many individuals with autism, leading to what is commonly described as “stimulus overselectivity,” “tunnel vision,” or “lack of multi-cue responsivity” (Lovaas et al. 1979, Ploog 2010, Schreibman 1988, Lovaas et al. 1971). Thirty to forty percent of individuals diagnosed with autism experience lifelong impairment in the ability to understand flexible syntax and spatial prepositions, and to mentally solve even the simplest of problems (Fombonne 2003). These individuals, commonly referred to having low-functioning ASD, typically exhibit full-scale IQ below 70 (Boucher et al. 2008, Beglinger and Smith 2001) and perform below the score of 85 in non-verbal IQ tests (see Boucher et al. 2008, table 14.1, Performance Quotient). Since significant language delay is a common problem in children with autism, lack of childhood use of syntactic language is likely one of the main contributing factors for this disability.

The period of plasticity for the development of synchronous frontoposterior connections seems to close in several stages. Based on the randomized controlled study of institutionalized Romanian children, the plasticity seems to already decline before the age of two: children placed in foster care and therefore exposed to complex syntactic language before the age of 2 showed increased myelination and performed better in mental integration tasks compared to children who have been placed in foster care after the age of 2 (Bick et al. 2015).

The next significant decline in plasticity is around the age of five. When the left hemisphere is surgically removed before the age of five (to treat cancer or epilepsy), patients often attain normal cognitive functions in adulthood (using the one remaining hemisphere). Removal of the left hemisphere after the age of five often results in significant impairment of syntactic language and tasks requiring mental synthesis (Basser 1962, Boatman et al. 1999, Krashen and Harshman 1972, Lenneberg 1967, Pulsifer et al. 2004). The decline of plasticity after the age of 5 likely corresponds to a consolidation of asymmetry between hemispheres with one of the hemispheres (usually the left one) taking most of the control of the language function with underlying left LPFC puppeteer synchronously connected to the rest of the cortex.

Further reduction of plasticity occurs by the time of puberty; a lack of experience in syntactic speech before puberty likely results in a consolidation of widely desynchronized frontoposterior networks and prevents individuals from subsequent acquisition of mental synthesis. While parts of the PFC network retain some plasticity for a significantly longer period of time, since myelination of the PFC continues into the third decade of life or later (Toga et al. 2006), this plasticity seems inadequate to assist patients who were linguistically deprived until puberty in the acquisition of mental synthesis despite many years of intensive therapy.

It is common among linguists to discuss the “language module” in the brain. In support of the existence of the “language module,” linguists enlist patients with apparent dissociation between language and cognition (Curtiss 1981). Observations in low functioning patients with fluid language production are pitted against the observations in high functioning individuals with aphasia. Is there real neurological dissociation between language and cognition and an easily separable language module? The answer to these questions depends on the definition of language.

The word “language” has highly divergent meanings in different contexts and disciplines. In informal usage, a language is usually understood as a culturally specific communication system (German, French, or Spanish Language), i.e. understanding of individual words without any reference to understanding of flexible syntax. On the other hand, in linguistics, the term “language” is used quite differently to encompass both understanding of words and flexible syntax. Fitch, Hauser, & Chomsky (Fitch et al. 2005) explain the ambiguity of the “language” definition: “It rapidly became clear in the conversations leading up to HCF (i.e. Hauser et al. 2002) that considerable confusion has resulted from the use of “language” to mean different things. We realized that positions that seemed absurd and incomprehensible, and chasms that seemed unbridgeable, were rendered quite manageable once the misunderstandings were cleared up. For many linguists, “language” delineates an abstract core of computational operations, central to language and probably unique to humans. For many biologists and psychologists, “language” has much more general and various meanings, roughly captured by “the communication system used by human beings.” Neither of these explananda are more correct or proper, but statements about one of them may be completely inapplicable to the other.”

Mental synthesis mediated by the LPFC is the best candidate for the “abstract core of computational operations, central to language and probably unique to humans” described by Fitch, Hauser, & Chomsky (Fitch et al. 2005). Mental synthesis is an essential part of all human languages. When we speak we use mental synthesis to describe a novel image (“My house is the second one on the left, just across the road from the red gate”) and we rely on the listener to use mental synthesis in order to visualize the novel image. When we tell stories, we are often describing things that the listener has never seen before (“That creature has two heads, three tails, dark eyes, and can run faster than a cheetah”) and we rely on the listener to imagine the story in their mind’s eye. As Steven Pinker put it, “the speaker has a thought, makes a sound, and counts on the listener to hear the sound and to recover that thought” (Pinker 2015).

All human languages are recursive: they allow high fidelity transmission of infinite number of novel images with the use of a finite number of words, Fig. 5. This is only possible because both listeners and speakers have mental synthesis ability presumably mediated by synchronous frontoposterior connections. Thus, both syntactic language and cognition depend on mental synthesis and on synchronous frontoposterior connections. The dissociation observed in many patients (Curtiss 1981) is not between syntactic language and cognition, but rather between their use of words (expressive language) and their mental synthesis ability often assessed by measuring their nonverbal IQ (Vyshedskiy et al. 2017). In fact, expressive language outside of a complex dialog cannot reliably serve as an indicator of mental synthesis ability. The following examples illustrate the problem of expressive language interpretation:

1) Some expressive language is completely automated, for example, coprolalia sometimes observed in patients with Tourette syndrome involves spontaneous and articulate uttering of expletives in an uncontrollable manner (Daniel and Perry 2016).

2) Tape recorder language and reading poetry by heart can sometimes be completely automated, prerecorded through repetition (Curtiss 1981, Zeman et al. 2013).

3) Echolalia in its profound form is automatic and effortless, imitative behavior whereby sounds or actions are imitated without explicit awareness (Mergl and Azoni 2015, Neely et al. 2016).

4) Confabulations involve the production of fabricated memories without the conscious intention to deceive. A liar would commonly initially simulate a story by using LPFC-controlled mental synthesis of multiple objects and check it for consistency before delivering the story to the listeners. The confabulator, by definition, does not have a pre-simulated story that could indicate the ability of the LPFC to ascertain control over the posterior cortex. Confabulatory tendencies are often interpreted as an indicator of a disconnection between PFC and the posterior cortex (Venneri et al. 2017).

5) Many people report the experience of reading aloud a book to a child and thinking of something else. In this case their LPFC is likely driving their mental synthesis while the book’s sensory input is driving the Broca’s area speech production and the right hemisphere's prosody in the bottom-up manner.

6) The vivid, original, bizarre, emotional, and story-like dreams that most people remember when they wake up occur during rapid eye movement (REM) sleep. These dreams consist of existing fragments of memory put into various new combinations and are often erroneously assumed to be created by the LPFC via the mechanism similar to mental synthesis. For example, Darwin associates imagination with dreaming: “The imagination is one of the highest prerogatives of man. By this faculty he unites former images and ideas, independently of the will, and thus creates brilliant and novel results... Dreaming gives us the best notion of this power... The value of the products of our imagination depends of course on the number, accuracy, and clearness of our impressions, on our judgment and taste in selecting or rejecting the involuntary combinations, and to a certain extent on our power of voluntarily combining them. As dogs, cats, horses, and probably all the higher animals, even birds have vivid dreams, and this is shown by their movements and the sounds uttered, we must admit that they possess some power of imagination.” (Darwin 1888).

Two lines of evidence point toward the mechanism of dreams that does not depend on the LPFC. First, neuroimaging observations indicate that the LPFC is inactive during sleep, both slow-wave sleep and REM sleep (Braun et al. 1997). Second, in people whose LPFC gets damaged, dreams do not change at all (Solms 1997). In dreams and hallucinations, images of people and objects may arise spontaneously and even hybridize into novel mental images. A subject can then describe these novel images and nonpresent people verbally, all without ever relying on LPFC-controlled mental synthesis.

We conclude that it is often impossible for an external observer to reliably infer internal imagery of the speaker based on expressive language. The speech may be automatic and independent of internal imagery, triggered spontaneously or driven by sensory input. Even when words reflect internal images, they may not inform of the LPFC control over imagination since even the novel images may have formed spontaneously as in dreaming or hallucination. There is no accurate method for external assessment of mental synthesis with respect to expressive language. While understanding of flexible syntax commonly requires LPFC-controlled mental synthesis, expressive language does not.

Several other bottom-up processes are often erroneously assumed to be indicative of the LPFC-controlled mental synthesis:

1) Object recognition / perceptual closure. Recognition of pictures with incomplete visual information is often erroneously assumed to be under the control of the LPFC. While there is no doubt that objects encoded in the posterior cortex can be primed by a category encoded in the PFC, there is strong evidence of bottom-up-driven PFC-independent process in perceptual closure tests, such as Moony faces test (Grützner et al. 2010). Accordingly, a good performance at these tests has no bearing on the LPFC-controlled mental synthesis abilities.

2) Drawing interpretation. The tendency to perceive objects from lines and colors has been used extensively in psychological tests such as blot test. It is often as easy to recognize an object from meaningless colors or scribbles of lines as it is from a cloud in the sky. Furthermore, once we recognize an object in a drawing, we tend to presume knowing the inclinations of the artist’s mind. However, the artist may have produced the drawing automatically or just spilled some colors as in blot test pictures.

Even some representational drawings can be fully automatic. For example, instrumental training (a variation of Pavlovian training) has been used to train elephants to produce representational works of art. Patients with visual agnosia can copy a drawing without any awareness of the type of object in the picture. The famous patient known in literature as pilot John faithfully copied a representation of St. Paul Cathedral without knowing the object in the picture (Humphreys and Riddoch 1987). Scribbling can often produce recognizable forms with no intention on the part of the scribbler. These examples are reminiscent of the problems with interpretation of expressive language. Outside of a complex dialog, neither expressive language, nor object recognition or drawing can inform on the speaker or artist’s LPFC-controlled mental synthesis ability. Only a complex syntactic dialog or representational drawing directed by a verbal description of a novel object (“a five headed horse”) can be a reliable indicator of the LPFC ability to combine disparate neuronal ensembles.