A brief outline of how the somatosensory system develops and what studies of developmental neurobiology can bring to bear on the questions of the pain experience as a fetus.

The ‘feeling’ of pain is simultaneously an intensely personal and yet universal part of human existence. Each of us, with few exceptions, will know when something is painful. It is a sense that is so fundamental to our conception of the world around us that most would struggle to imagine a life without pain. Our pain ‘faculties’, however, have to come from somewhere. The ability to appreciate the sensuousness of silk or the irritating sting of a papercut derives from the intricate neural circuitry of the somatosensory system. With a growing number of extremely preterm infants enduring prolonged stays in intensive care, the need to understand how and when the ability to process pain develops in the fetal life has become more pressing.

The International Association for the Study of Pain (IASP) defines pain as, ‘an unpleasant sensory and emotional experience associated with actual or potential tissue damage, or described in terms of such damage.’ IASP goes on to say that the, ‘inability to communicate verbally does not negate the possibility that an individual is experiencing pain and is in need of appropriate pain-relieving treatment.’ In practice, therefore, whether the patient is a fully conscious adult, premature neonate or fetus, there is an obligation to assess and treat pain where it exists.


Much of what is understood about normal human neural development has arisen from studies of the developing nervous systems of flies, frogs and rodents. Investigators have mapped out the origins and fate of immature neurones from the primitive neural plate to the adult spinal cord to better understand the complex web of cellular relationships that exist throughout our nervous tissue. Yet, in isolation, a single neurone cannot conduct the veritable symphony of activity required for the sensation of pain. From skin to cortex, signals conveying sensory information travel through multiple synapses. The development of these connections is outlined below.

In the periphery, a noxious stimulus is transduced into action potentials in the terminals of primary sensory neurones (free nerve endings) – present in the skin from approximately seven to 15 weeks gestation. These signals then pass through the dorsal root ganglion and on to the spinal cord, where they synapse on numerous neurones. This spinal cord circuit is an important site of integration and modulation of sensory signals before they are transmitted up to the brainstem and thalamus, which are further key supraspinal stations for control and distribution of sensory information. Output from these subcortical areas is then transmitted to the cortex. Thalamo-cortical connections are first observed in a region just below the primitive cortex known as the subplate between 20 and 22 weeks and soon move into the immature cortex between 23 and 24 weeks.1

‘Feeling’ pain, in adults at least, is not simply the result of activation in the somatosensory cortex. Numerous parts of the brain are involved in the full experience of pain. In addition to the salience, localisation and sensation of the noxious stimulus, there is an emotional component that, in the adult, is fundamental to the concept of pain. Together, these processes require both thalamo-cortical and intra-cortical connections. Current scientific evidence suggests that a functional cortex is a basic requirement for pain experience in the fetus.2

Fetal reflexes and hormonal stress responses in early gestation have supported the hypothesis that pain may be felt before 24 weeks. A fetus is observed to withdraw from surgical tools in a reflex manner. Similarly, studies have shown that when transfusing the intra-hepatic vein a substantial stress response is mounted by the fetus.3,4 Beta-endorphin, an endogenous opioid, is released in response to needling as early as 18 weeks and cortisol responses to the procedure can be seen at 20 weeks. Additionally, stimulation can result in an increased heart rate and diversion of blood to the core away from the periphery accompanied by changes in plasma glucose, adrenaline and noradrenaline. Proponents of the idea that the fetus is sensitive earlier in development point to studies showing that stress responses can be attenuated with the use of fentanyl, an analgesic, in combination with a neuromuscular blocking agent and nitrous oxide.

Such responses, however, occur prior to the development of functional thalamo-cortical connections and, therefore, are likely to be mediated at the subcortical level. Movements in surgery can be fully explained by the presence of intact reflex arcs through the spinal cord, which develop between seven-and-a-half and 14 weeks, though specific responses to nociceptive stimuli do not appear until 19 weeks. Stress responses do not provide direct evidence of fetal pain. Instead, they are a demonstration of the fetus’s neural response to threatening stimuli.

If we accept – in light of the neuroanatomical evidence above – that connections between the thalamus and the cortex are necessary for pain, then the fetus is unlikely to experience pain before 24 weeks.


Whether the fetus is conscious and therefore aware of pain is another factor that needs to be examined when determining the presence or absence of fetal ‘feeling’. If consciousness is taken to mean a state of wakeful awareness, then one needs to establish that the fetus is both awake and aware of its surroundings. Being awake is not, in itself, a sign of consciousness, though one must be awake in order to be conscious.5

Wakefulness is coordinated at the level of the brainstem and, as such, requires no cortical activity. It is difficult to determine whether the fetus is in fact ever awake while in the womb. Through the use of EEG in the sheep fetus, investigators have shown evidence for the existence of two distinct sleep states, namely rapid-eye-movement (REM) sleep and non-REM (NREM) sleep. These are most developed in late gestation and are the forerunners of sleep states in the newborn. During REM sleep the fetus can be seen to make characteristic movements of the eyes, tongue and respiratory muscles; whereas, in contrast, there is substantially less movement in NREM sleep. Importantly, it is estimated that, as a fetus nears term, these sleep states account for approximately 95 per cent of EEG activity. In a recent review of the evidence of fetal awareness, Mellor and colleagues concluded ‘there is no strong evidence that the fetus is awake, even transiently.’5 In short, the absence of wakefulness precludes consciousness in the fetus and, by extension, fetal pain.

Though it is understood that the fetus is unlikely to be aware in the womb, this does not mean that noxious stimulation in utero can be ignored. The central and peripheral nervous systems undergo profound changes throughout gestation and the normal maturation of the cortex relies, in part, on the signals it receives from the outside world and from movement in the womb. Cushioned by amniotic fluid and potentially kept in a sleep-like state, the fetus’s exposure to noxious stimulation is very limited. In the case of extremely premature infants, as young as 23 weeks, the immature nervous system is thrown into a world of sensory information. Due to their immaturity, these infants are committed to prolonged stays in intensive care units. Within this setting, pain is a daily reality for these patients. Frequently performed procedures, such as heel lancing, intubation and cannulation, are both invasive and painful. The immaturity of the cortex brings the potential for cortical reorganisation and spinal sensitisation due to noxious input. As a result, clinicians have turned to analgesics to prevent the adverse consequences of painful interventions.

The mainstay of pain relief in neonates has been oral sucrose solution, given before a procedure. Sucrose has been convincingly shown in numerous studies to reduce facial (nasolabial furrow, eye squeeze and brow bulge) and physiological (heart rate, oxygen saturation) scores of pain – widely used to bridge the communication divide between physicians and babies. Interestingly, however, our group has recently shown that cortical activation, measured by EEG, following clinically required noxious heel lance, is not significantly different from that in premature infants given water instead of sucrose.6 Thus, it would seem that the risk of exposure to repeated and painful interventions may remain in the presence of sucrose and we need to undertake further research into other, more effective forms of pain relief.7


As with most individuals who cannot communicate, it may never be possible to categorically determine when exactly the human fetus or neonate feels pain. The evidence for the existence of fetal pain is weak. Hormonal responses reflect the activity of areas other than the cortex and act as surrogate measures of arousal rather than pain. The neural building blocks that relay nerve impulses from the skin to the cortex are likely to be in place by at least 24 weeks gestation, but the intracortical connections required for the full pain experience take longer to become established. In addition, there is evidence that the fetus may not be conscious and therefore have no awareness during this time.

What cannot be ignored, however, is the profound ability of sensory information to shape the developing brain. Adverse noxious events both within the womb and beyond, in neonatal intensive care, may have detrimental effects upon the developing nervous system. Thus, effective analgesia in very immature infants could be viewed more as relief from future problems rather than from immediate pain.